Rails - Aug 2005 - Pluralized Controller Names?

If I do:

    script/generate controller Photo

I get a controller named photo_controller. OTOH, if I do:

     script/generate scaffold Photo

I get a controller named photos_controller. (Note that the controller  
name is pluralized).

I realize that I''m specifying the controller name explicitly in the  
first case, but it seems
odd to me that the scaffold command generates a pluralized modelname  
in the absence
of a specific controller name. It also seems odd that there appears  
to be an inconsistency
here: why pluralize in one case, but not the other?

Does anybody know why the controller name gets pluralized if you  
don''t specify it explictly
when generating scaffolding? Also, from what I can tell, controller  
names, by convention,
are not typically pluralized, so I find it curious that Rails  
pluralizes the controller name in this case.


david

On 8/8/05, David Geary <sabreware-ihVZJaRskl1bRRN4PJnoQQ@public.gmane.org>
wrote:
> If I do:
> 
>     script/generate controller Photo
> 
> I get a controller named photo_controller. OTOH, if I do:
> 
>      script/generate scaffold Photo
> 
> I get a controller named photos_controller. (Note that the controller
> name is pluralized).
> 
> I realize that I''m specifying the controller name explicitly in
the
> first case, but it seems
> odd to me that the scaffold command generates a pluralized modelname
> in the absence
> of a specific controller name. It also seems odd that there appears
> to be an inconsistency
> here: why pluralize in one case, but not the other?
In the second case (script/generate scaffold) you''ve specified the
name of the Model, and left the name of the controller for Rails to
guess.  It will guess the name to be the pluralization of the model
name.

You might want ''script/generate scaffold Photo photo''

Or just run ''script/generate scaffold'' to get the
documentation.

> 
> Does anybody know why the controller name gets pluralized if you
> don''t specify it explictly
> when generating scaffolding? Also, from what I can tell, controller
> names, by convention,
> are not typically pluralized, so I find it curious that Rails
> pluralizes the controller name in this case.
Convention is that Model name is singular (Photo) and Controller name
is plural (photos_controller).  Now many people do prefer singular
controller names, but they can argue that point with DHH...


Jason

Le 05-08-08 à 15:04, Jason Foreman a écrit :
> On 8/8/05, David Geary
<sabreware-ihVZJaRskl1bRRN4PJnoQQ@public.gmane.org> wrote:
>
>> If I do:
>>
>>     script/generate controller Photo
>>
>> I get a controller named photo_controller. OTOH, if I do:
>>
>>      script/generate scaffold Photo
>>
>> I get a controller named photos_controller. (Note that the controller
>> name is pluralized).
>>
>> I realize that I''m specifying the controller name explicitly
in the
>> first case, but it seems
>> odd to me that the scaffold command generates a pluralized modelname
>> in the absence
>> of a specific controller name. It also seems odd that there appears
>> to be an inconsistency
>> here: why pluralize in one case, but not the other?
>>
>
> In the second case (script/generate scaffold) you''ve specified the
> name of the Model, and left the name of the controller for Rails to
> guess.  It will guess the name to be the pluralization of the model
> name.
> You might want ''script/generate scaffold Photo photo''
>
> Or just run ''script/generate scaffold'' to get the
documentation.
Thanks, but I knew that already. 8-)
>
>> Does anybody know why the controller name gets pluralized if you
>> don''t specify it explictly
>> when generating scaffolding? Also, from what I can tell, controller
>> names, by convention,
>> are not typically pluralized, so I find it curious that Rails
>> pluralizes the controller name in this case.
>>
>
> Convention is that Model name is singular (Photo) and Controller name
> is plural (photos_controller).
Yup, I knew that too. I''m trying to understand why it''s that
way.
>   Now many people do prefer singular
> controller names, but they can argue that point with DHH...
I don''t know if that''d do any good. The pragmatic book uses
singular
controller
names and DHH''s name is on the cover!


david
>
>
> Jason
> _______________________________________________
> Rails mailing list
> Rails-1W37MKcQCpIf0INCOvqR/iCwEArCW2h5@public.gmane.org
> http://lists.rubyonrails.org/mailman/listinfo/rails
>

On 8/8/05, David Geary <sabreware-ihVZJaRskl1bRRN4PJnoQQ@public.gmane.org>
wrote:
> Le 05-08-08 à 15:04, Jason Foreman a écrit :
> 
> > On 8/8/05, David Geary
<sabreware-ihVZJaRskl1bRRN4PJnoQQ@public.gmane.org> wrote:
> >
> >> If I do:
> >>
> >>     script/generate controller Photo
> >>
> >> I get a controller named photo_controller. OTOH, if I do:
> >>
> >>      script/generate scaffold Photo
> >>
> >> I get a controller named photos_controller. (Note that the
controller
> >> name is pluralized).
> >>
> >> I realize that I''m specifying the controller name
explicitly in the
> >> first case, but it seems
> >> odd to me that the scaffold command generates a pluralized
modelname
> >> in the absence
> >> of a specific controller name. It also seems odd that there
appears
> >> to be an inconsistency
> >> here: why pluralize in one case, but not the other?
> >>
> >
> > In the second case (script/generate scaffold) you''ve
specified the
> > name of the Model, and left the name of the controller for Rails to
> > guess.  It will guess the name to be the pluralization of the model
> > name.
> > You might want ''script/generate scaffold Photo
photo''
> >
> > Or just run ''script/generate scaffold'' to get the
documentation.
> 
> Thanks, but I knew that already. 8-)
> 
> >
> >> Does anybody know why the controller name gets pluralized if you
> >> don''t specify it explictly
> >> when generating scaffolding? Also, from what I can tell,
controller
> >> names, by convention,
> >> are not typically pluralized, so I find it curious that Rails
> >> pluralizes the controller name in this case.
> >>
> >
> > Convention is that Model name is singular (Photo) and Controller name
> > is plural (photos_controller).
> 
> Yup, I knew that too. I''m trying to understand why it''s
that way.
Oh, I thought your question was more technical, as in why that command
generated a pluralized controller, as opposed to a philosophical
question on why Rails uses plural controller names.  My apologies :)
> 
> >   Now many people do prefer singular
> > controller names, but they can argue that point with DHH...
> 
> I don''t know if that''d do any good. The pragmatic book
uses singular
> controller
> names and DHH''s name is on the cover!
Well I guess I should confess to having bought the book but not yet
actually reading it...

DHH, you got some ''splainin to do!!  I wonder if that might be more
the work of Dave Thomas than DHH though?

I personally like the scheme of singular model, plural controller. 
The model is sort of defining the characteristics of the instances of
the model.  So a ''Photo'' has one ''Artist''. 
But the controller
typically has functions to manipulate the ''Photos'', ie the
controller
of the Photos, or the photos_controller.


Sorry I couldn''t be more help!

Jason

Le 05-08-08 à 15:29, Jason Foreman a écrit :
> On 8/8/05, David Geary
<sabreware-ihVZJaRskl1bRRN4PJnoQQ@public.gmane.org> wrote:
>
>> Le 05-08-08 à 15:04, Jason Foreman a écrit :
>>
>>
>>> On 8/8/05, David Geary
<sabreware-ihVZJaRskl1bRRN4PJnoQQ@public.gmane.org> wrote:
>>>
>>>
>>>> If I do:
>>>>
>>>>     script/generate controller Photo
>>>>
>>>> I get a controller named photo_controller. OTOH, if I do:
>>>>
>>>>      script/generate scaffold Photo
>>>>
>>>> I get a controller named photos_controller. (Note that the  
>>>> controller
>>>> name is pluralized).
>>>>
>>>> I realize that I''m specifying the controller name
explicitly in the
>>>> first case, but it seems
>>>> odd to me that the scaffold command generates a pluralized  
>>>> modelname
>>>> in the absence
>>>> of a specific controller name. It also seems odd that there
appears
>>>> to be an inconsistency
>>>> here: why pluralize in one case, but not the other?
>>>>
>>>>
>>>
>>> In the second case (script/generate scaffold) you''ve
specified the
>>> name of the Model, and left the name of the controller for Rails to
>>> guess.  It will guess the name to be the pluralization of the model
>>> name.
>>> You might want ''script/generate scaffold Photo
photo''
>>>
>>> Or just run ''script/generate scaffold'' to get the
documentation.
>>>
>>
>> Thanks, but I knew that already. 8-)
>>
>>
>>>
>>>
>>>> Does anybody know why the controller name gets pluralized if
you
>>>> don''t specify it explictly
>>>> when generating scaffolding? Also, from what I can tell,
controller
>>>> names, by convention,
>>>> are not typically pluralized, so I find it curious that Rails
>>>> pluralizes the controller name in this case.
>>>>
>>>>
>>>
>>> Convention is that Model name is singular (Photo) and Controller  
>>> name
>>> is plural (photos_controller).
>>>
>>
>> Yup, I knew that too. I''m trying to understand why
it''s that way.
>>
>
> Oh, I thought your question was more technical, as in why that command
> generated a pluralized controller, as opposed to a philosophical
> question on why Rails uses plural controller names.  My apologies :)
>
>
>>
>>
>>>   Now many people do prefer singular
>>> controller names, but they can argue that point with DHH...
>>>
>>
>> I don''t know if that''d do any good. The pragmatic
book uses singular
>> controller
>> names and DHH''s name is on the cover!
>>
>
> Well I guess I should confess to having bought the book but not yet
> actually reading it...
>
> DHH, you got some ''splainin to do!!  I wonder if that might be
more
> the work of Dave Thomas than DHH though?
> I personally like the scheme of singular model, plural controller.
> The model is sort of defining the characteristics of the instances of
> the model.  So a ''Photo'' has one
''Artist''.  But the controller
> typically has functions to manipulate the ''Photos'', ie
the controller
> of the Photos, or the photos_controller.
Yes, I agree completely. However, in the book, singular names make  
sense; for example Admin. I
don''t think you''d want a plural name in that case.

Perhaps the answer is that when you use the same name for both
model and controller (by not specifying a controller name explicitly)
it makes sense to pluralize the controller name. If you have a specific
name, you might not want it to be pluralized (like Admin, for  
example). And,
perhaps, I''ve just answered my own question. 8-)
> Sorry I couldn''t be more help!
Hey, no problem. I appreciate your replies.


david
>
> Jason
> _______________________________________________
> Rails mailing list
> Rails-1W37MKcQCpIf0INCOvqR/iCwEArCW2h5@public.gmane.org
> http://lists.rubyonrails.org/mailman/listinfo/rails
>

On Aug 8, 2005, at 5:14 PM, David Geary wrote:
> Le 05-08-08 à 15:04, Jason Foreman a écrit :
>
>>   Now many people do prefer singular
>> controller names, but they can argue that point with DHH...
>>
>
> I don''t know if that''d do any good. The pragmatic book
uses
> singular controller
> names and DHH''s name is on the cover!
It needs to be said: Auto pluralization and singularization are evil.  
If you disagree, please look at this patch:

<http://dev.rubyonrails.com/attachment/ticket/1844/ 
big_inflector_fix.patch>

Did you know "schemata" is the plural of "schema"? And that
if you
name your class "Virus", Rails will look for the table named
"viri"?
The two dictionaries I consulted seem to think the plural is  
"viruses". Stupid dictionaries!

Regards,
Ed

-- 
Transmogrify, LLC * <http://xmog.com/>

> It needs to be said: Auto pluralization and singularization are evil.
> If you disagree, please look at this patch:
Also look at all the other patches that are there around
inflector.pluralize. Various words have two plurals, equipment, index
being two common examples.  Money being another (sums of money
pluralises to monies)

Even leaving aside all the new words which the english language
absorbs every year,   there simply is no way to solve this problem
reliably.

Having said all that,   it is what it is, and we can''t just break
every application by turning off pluralisation by default.


-- 
Cheers

Koz

>
> It needs to be said: Auto pluralization and singularization are evil.  
> If you disagree, please look at this patch:
>
> <http://dev.rubyonrails.com/attachment/ticket/1844/ 
> big_inflector_fix.patch>
>
> Did you know "schemata" is the plural of "schema"? And
that if you
> name your class "Virus", Rails will look for the table named
"viri"?
> The two dictionaries I consulted seem to think the plural is  
> "viruses". Stupid dictionaries!
Well I actually wrote that patch to address a few bugs that were on the 
buglist - so do you address the bugs that people post or not?  I thought 
it''d be worthwhile fixing some/most of the requests, but it''s
not up to
me to actually include that code in the rails code-base.  This is how 
open source works (at least in all the projects I''ve participated in), 
people contribute, and a committer decides if the contribution addresses 
a need or fixes an outstanding bug.

As for Virus/Viri.  The fact is that virus does not have a recognized 
plural form in it''s original language (I think Latin, but I''m
not
positive).  Viruses is the ''anglicized'' version.  Rails (at
least my
local version :) now uses ''viruses'' as the plural of virus. 
If you want
to know more about algorithmic pluralization/singularization, then there 
are papers written on the topic and there is indeed a perl module that 
encapsulates the rules (we have at least 50% of these in the inflector 
right now, I''d actually say that in the patch we get close to 75%, 
catching the others is difficult as it depends on context).  Schema -> 
Schemata, is very very odd, but then what is the plural of Octopus?  
It''s actually Octopodes (not Octopi - I have to shout at my old biology
teacher about not knowing that one).  So should we use something that is 
correct in terms of the roots of the language (Octopus come from greek I 
think), or should we go with common usage?  Of course the simple answer 
is to do zero pluralization and then everything is fine (or broken, but 
all in the same way).

In summary, it is a problem, but one that has an algorithmic solution, 
so the questions are,
do we address the problem?
do we use common usage as a guide,
when we can make a fix to resolve someones problem do we do it?,
do we aim for 100% accuracy (as the perl solution attempts),
do we address the problems as the appear? (seemingly the current strategy),
do we provide no pluralization (which people new to rails actually think 
is pretty cool in my experience)

But your opinion may be different.

Kev

Perhaps one way to make things a bit more straightforward would be to
have a web-based plural lookup. So you type in "octopus" and it
returns the pluralisation that Rails knows...

Of course, most are obvious, but it could be helpful for some of the
more complex ones and especially for newbies. Maybe even a simple Ruby
script in the script directory would be better?

-- 
Ben Myles
railsapphosting.com

On 8/9/05, Kev Jackson
<kevin.jackson-1n8Jz8fJjQ+aMJb+Lgu22Q@public.gmane.org> wrote:
...
> In summary, it is a problem, but one that has an algorithmic solution,
> so the questions are,
> do we address the problem?
> do we use common usage as a guide,
> when we can make a fix to resolve someones problem do we do it?,
> do we aim for 100% accuracy (as the perl solution attempts),
> do we address the problems as the appear? (seemingly the current strategy),
> do we provide no pluralization (which people new to rails actually think
> is pretty cool in my experience)

On 8/8/05, Ben Myles <ben.myles-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> Perhaps one way to make things a bit more straightforward would be to
> have a web-based plural lookup. So you type in "octopus" and it
> returns the pluralisation that Rails knows...
Geoff Grosenbach has provided us with exactly that:

http://nubyonrails.topfunky.com/articles/2005/08/04/the-amazing-rails-pluralizer

Assuming that''s what you meant, of course. :)

-- 
Chris Boone

http://hypsometry.com/  :  website edification
http://uvlist.org/  :  free classifieds for the Upper Valley

Very cool!! That''s exactly what I meant :-)

On 8/9/05, Chris Boone
<hypsometry-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> Geoff Grosenbach has provided us with exactly that:
> 
>
http://nubyonrails.topfunky.com/articles/2005/08/04/the-amazing-rails-pluralizer
...

Ben Myles wrote:
>Perhaps one way to make things a bit more straightforward would be to
>have a web-based plural lookup. So you type in "octopus" and it
>returns the pluralisation that Rails knows...
>
>  
>
There is one

http://nubyonrails.topfunky.com/ uses rails/inflector to return plurals

Kev

Hallelujah.

I think that in a community where "convention over configuration" is  
espoused as a godsend, we should really get together and establish  
which conventions are necessities and which are problematic,  
superfluous, etc. As great as the marketing and branding efforts for  
RoR have been thus far, the more people dig deeper or try to apply  
the framework to existing apps, the more backlash there is gonna be  
over things like pluralization. So many recorded and searchable  
claims on blogs and on this list in the fashion of "if it can''t be
done in Rails it isn''t worth doing" are not going to help the PR  
effort when departments start arguing over the adoption of Ruby and  
Rails.

Turning off pluralization in updates may or may not be acceptable at  
this point -- the user base is still small enough that it may be  
prime time to do it -- but there could be a deprecation period and an  
AR mix-in for future use. Though, a blanket update *could* be the  
best option, as long as the update is mentioned in release notes.  
After all - expecting problems when blindly updating software on  
production systems seems obvious enough to be convention to me... ;)







On Aug 8, 2005, at 6:45 PM, Ed Watkeys wrote:
> It needs to be said: Auto pluralization and singularization are evil.


_______________________________________________
Rails mailing list
Rails-1W37MKcQCpIf0INCOvqR/iCwEArCW2h5@public.gmane.org
http://lists.rubyonrails.org/mailman/listinfo/rails

> Turning off pluralization in updates may or may not be acceptable at this
> point -- the user base is still small enough that it may be prime time to
do
> it -- but there could be a deprecation period and an AR mix-in for future
> use.
No,  I''m certain that you''ll find that *all* of the core team
have
rails applications which rely on pluralisation.   The odds are slim if
you expect commiters to break their own apps, and those of everyone
else.
> Though, a blanket update *could* be the best option, as long as the
> update is mentioned in release notes. After all - expecting problems when
> blindly updating software on production systems seems obvious enough to be
> convention to me... ;)
> 
> 
> 
> 
> 
> 
> 
> 
> On Aug 8, 2005, at 6:45 PM, Ed Watkeys wrote:
> 
> 
> It needs to be said: Auto pluralization and singularization are evil. 
>  
> _______________________________________________
> Rails mailing list
> Rails-1W37MKcQCpIf0INCOvqR/iCwEArCW2h5@public.gmane.org
> http://lists.rubyonrails.org/mailman/listinfo/rails
> 
> 
> 

-- 
Cheers

Koz

On 8/8/05, Ben Myles <ben.myles-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> Very cool!! That''s exactly what I meant :-)
Seems slightly broken, though.

Apparently you have two Octopus but I have only one Octopu.

-- 
Urban Artography
http://artography.ath.cx

On Aug 8, 2005, at 10:23 PM, Toby Boudreaux wrote:
> Hallelujah.
>
> I think that in a community where "convention over configuration"
> is espoused as a godsend, we should really get together and  
> establish which conventions are necessities and which are  
> problematic, superfluous, etc. As great as the marketing and  
> branding efforts for RoR have been thus far, the more people dig  
> deeper or try to apply the framework to existing apps, the more  
> backlash there is gonna be over things like pluralization. So many  
> recorded and searchable claims on blogs and on this list in the  
> fashion of "if it can''t be done in Rails it isn''t
worth doing" are
> not going to help the PR effort when departments start arguing over  
> the adoption of Ruby and Rails.
>
> Turning off pluralization in updates may or may not be acceptable  
> at this point -- the user base is still small enough that it may be  
> prime time to do it -- but there could be a deprecation period and  
> an AR mix-in for future use. Though, a blanket update *could* be  
> the best option, as long as the update is mentioned in release  
> notes. After all - expecting problems when blindly updating  
> software on production systems seems obvious enough to be  
> convention to me... ;)
There are a lot of things going on in Rails. Some involve convention.  
Others involve inference. I am not a big fan of software making  
inferences about the configuration of a web application based on the  
vagaries of English pluralization rules. Nor am I a big fan of a  
system''s out-of-the-box configuration being determined by what a  
newbie PHP refugee knows about the art and science of database modeling.

Regards,
Ed

-- 
Transmogrify, LLC * <http://xmog.com/>

Why can''t we have both pluralisation and non-pluralisation at the same
time? Ie. Rails first looks for a table in the plural form, if it''s
not found it looks for one that''s not a plural. This way it
shouldn''t
break existing code and it''s a really simple way to let people choose
the route they want to follow.

Ben

On Aug 8, 2005, at 10:32 PM, Michael Koziarski wrote:
>> Turning off pluralization in updates may or may not be acceptable  
>> at this
>> point -- the user base is still small enough that it may be prime  
>> time to do
>> it -- but there could be a deprecation period and an AR mix-in for  
>> future
>> use.
>>
>
> No,  I''m certain that you''ll find that *all* of the core
team have
> rails applications which rely on pluralisation.   The odds are slim if
> you expect commiters to break their own apps, and those of everyone
> else.
The goal is not to keep running the <1000 web apps that are currently  
built atop Rails, but to make Rails a compelling platform for a  
million more.

Regards,
Ed

-- 
Transmogrify, LLC * <http://xmog.com/>

I agree entirely. I think these rules should be determined by users  
during these initial "growing pains" in the relatively first days of  
popular adoption. I don''t know that there are any "newbie PHP  
refugees" trying to determine the configuration with Rails, but
I''m
sure configuration won''t be solely turned over to such folks anyway  
-- nor that they''d be making many suggestions in the first place.

> Nor am I a big fan of a system''s out-of-the-box configuration
being
> determined by what a newbie PHP refugee knows about the art and  
> science of database modeling.
>

> The goal is not to keep running the <1000 web apps that are currently
> built atop Rails, but to make Rails a compelling platform for a
> million more.
My goal is to have my applications keep working, the same can be said
for everyone else who uses rails.

If you want another million applications to be built with rails,  we
can''t just go breaking the installed base when we find something we
don''t like.   That is *much* more annoying (ask anyone who moved from
rails 0.8->0.13, things changed all over the place).   Just like we
continue to support the old style finders, rails will continue to
support pluralised table names, and apps which use them.

They key is to find a way to support pluralisation as ''off''
for new
applications, but not break existing applications when they try to
upgrade.  This is certainly possible.

-- 
Cheers

Koz

AFAIK, the original reason for auto pluralization was twofold:

1. Plural table names and singular database names make rails apps
easier and more natural to discuss conversationally -- it fits with
the way we think about them.
2. It''s way cooler to automatically determine the plural form of a
word than to specify it in every file (convention over configuration)

The reasoning behind auto pluralization is still valid, and it does
provide the benefit it claims to provide (whether you care about that
benefit is a different issue entirely).  Why not just leave things the
way they are right now, and just make sure newbies know how to
manually specify the table name of a model, for those cases where the
inflector falls short?

On 8/9/05, Ben Myles <ben.myles-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> Why can''t we have both pluralisation and non-pluralisation at the
same
> time? Ie. Rails first looks for a table in the plural form, if
it''s
> not found it looks for one that''s not a plural. This way it
shouldn''t
> break existing code and it''s a really simple way to let people
choose
> the route they want to follow.
This is backwards compatible,  so it''s more likely to get accepted. 
Care to implement a patch?


-- 
Cheers

Koz

consider apple''s one button mouse.

On 8/8/05, Michael Koziarski
<koziarski-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> On 8/9/05, Ben Myles
<ben.myles-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> > Why can''t we have both pluralisation and non-pluralisation at
the same
> > time? Ie. Rails first looks for a table in the plural form, if
it''s
> > not found it looks for one that''s not a plural. This way it
shouldn''t
> > break existing code and it''s a really simple way to let
people choose
> > the route they want to follow.
> 
> This is backwards compatible,  so it''s more likely to get
accepted.
> Care to implement a patch?
I think this is great idea.  Might even be nice to check for the
plural or singular version of missing models and controllers as well
as tables.  This would be a great help towards making people''s first
experiences with Rails more positive.  I know I had quite a bit of
head-scratching at first about why my code wasn''t working, when it was
looking for some things as singular as some as plural.

- Isaac

Well if none of the core developers have any gripes I''ll have a look
at implementing a patch. I can''t promise anything though as I
haven''t
studied the intimate internals of Rails yet. It should be a good
learning experience.

-- 
Ben Myles
railsapphosting.com

On 8/9/05, Isaac Reuben <ireuben-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> I think this is great idea.  Might even be nice to check for the
> plural or singular version of missing models and controllers as well
> as tables.  This would be a great help towards making people''s
first
> experiences with Rails more positive.  I know I had quite a bit of
> head-scratching at first about why my code wasn''t working, when it
was
> looking for some things as singular as some as plural.

Why not standardize on a single source for pluralization rules, such
as Merriam-Webster, (m-w.com) and then make it easy for people to
override (which it probably already is) the built in functionality for
personal preference and/or backwards compatability?

On 8/8/05, Isaac Reuben <ireuben-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> On 8/8/05, Michael Koziarski
<koziarski-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> > On 8/9/05, Ben Myles
<ben.myles-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> > > Why can''t we have both pluralisation and
non-pluralisation at the same
> > > time? Ie. Rails first looks for a table in the plural form, if
it''s
> > > not found it looks for one that''s not a plural. This way
it shouldn''t
> > > break existing code and it''s a really simple way to let
people choose
> > > the route they want to follow.
> >
> > This is backwards compatible,  so it''s more likely to get
accepted.
> > Care to implement a patch?
> 
> I think this is great idea.  Might even be nice to check for the
> plural or singular version of missing models and controllers as well
> as tables.  This would be a great help towards making people''s
first
> experiences with Rails more positive.  I know I had quite a bit of
> head-scratching at first about why my code wasn''t working, when it
was
> looking for some things as singular as some as plural.
> 
> - Isaac
> _______________________________________________
> Rails mailing list
> Rails-1W37MKcQCpIf0INCOvqR/iCwEArCW2h5@public.gmane.org
> http://lists.rubyonrails.org/mailman/listinfo/rails
>

On Aug 8, 2005, at 6:23 PM, Kev Jackson wrote:
> Schema -> Schemata, is very very odd, but then what is the plural  
> of Octopus?  It''s actually Octopodes (not Octopi - I have to shout
> at my old biology teacher about not knowing that one).
Actually, Octopodes is not the common -- or preferred -- form of the  
plural. The OED (normally very conservative in usage) lists  
Octopuses, Octopi, and Octopodes, in that order.

Octopuses is generally the preferred plural. Fowler (the pedant''s  
choice for usage) agrees.

http://en.wikipedia.org/wiki/Octopus

Go, pods, go!

>
> Actually, Octopodes is not the common -- or preferred -- form of the  
> plural. The OED (normally very conservative in usage) lists  
> Octopuses, Octopi, and Octopodes, in that order.
>
The point I was trying to make was do we try to use common usage as a 
guide, or the real genuine historical plural - Octopodes is not common, 
but it is the correct one - all the others are newer/incorrect (for a 
given value of correct) - see 
http://www.csse.monash.edu.au/~damian/papers/HTML/Plurals.html for more info

Actually at the moment octopus is probably pluralised as octopi or 
octopuses in rails
> Octopuses is generally the preferred plural. Fowler (the pedant''s
> choice for usage) agrees.

Who is fowler?  I immediately thought of Martin Fowler, but I doubt he 
cares a jot about linguistic nuances

Kev

On Aug 8, 2005, at 11:16 PM, Michael Koziarski wrote:
>> The goal is not to keep running the <1000 web apps that are
currently
>> built atop Rails, but to make Rails a compelling platform for a
>> million more.
>>
>
> My goal is to have my applications keep working, the same can be said
> for everyone else who uses rails.
While in this case, breaking people''s code to make such a change is  
unnecessary and unadvised, I disagree with your assertion, or at  
least your assertion''s implication, that everyone''s goal is to
have
Rails never break their code. A web app is a means to an end, and if  
breaking some code makes Rails more suited to achieving a people''s  
ultimate goals -- making money, curing cancer, distributing warez --  
they SHOULD be happy to upgrade and make whatever changes are  
necessary. It''s not as if anyone is forcing them to upgrade
immediately.

Deciding whether to break code is complex. Keeping a deprecated  
interface around makes life easier for developers but delays the  
inevitable and makes the system more complex. Complex means bigger  
and slower. There''s a tradeoff. This early in the life of Rails, I  
would hope the Rails core developers have an aggressive approach to  
purging deprecated code from the system.

A project can get popular too quickly. Popularity means an installed  
base means people how don''t want their code to break. You need just  
the right adoption curve. The hype surrounding Rails perhaps  
attracted too many mainstream users when Rails really needs thrill- 
seeking, risk-tolerant early adopters.

Regards,
Ed

-- 
Transmogrify, LLC * <http://xmog.com/>

I may be asking for it, but why even bother with auto-pluralization? 
It seems like an unnecessary gimmick, in contrast to the Rails
philosophy of keeping things simple.  And not only does it add
complexity that''s (imo) not necessary, it adds confusion in some
cases.  I''m just guessing, but I imagine ESL speaking people might
have even more problems early on getting used to it.

Why is the system necessary?  Why do table names need to be plural?

For example, before Rails I would call my table "person", because it
defines a person. Yes it hold a collection of persons, but the table
definition describes a person.  I''ve always been comfortable saying
"select * from person where lastname like ''Thom%''.  To me
it''s the
"person table".

On 8/8/05, Ben Myles <ben.myles-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> Why can''t we have both pluralisation and non-pluralisation at the
same
> time? Ie. Rails first looks for a table in the plural form, if
it''s
> not found it looks for one that''s not a plural. This way it
shouldn''t
> break existing code and it''s a really simple way to let people
choose
> the route they want to follow.
This is going about it backwards. The point of getting rid of
pluralisation is so that we don''t have to waste processing power
computing the plural of the word. So the code should by default check
for the *singular* table name, and if it''s not found, *then* fall back
to the pluralised table name.

-- 
Urban Artography
http://artography.ath.cx

> Why is the system necessary?  Why do table names need to be plural?
> 
> For example, before Rails I would call my table "person", because
it
> defines a person. Yes it hold a collection of persons, but the table
> definition describes a person.  I''ve always been comfortable
saying
> "select * from person where lastname like ''Thom%''. 
To me it''s the
> "person table".
I think it''s because it''s what David prefers, and it works
with most
cases.  But, it''s not necessary by any means.  Just add this to your
environment.rb:

ActiveRecord::Base.pluralize_table_names = false 

If you have a beef with the id primary keys and would prefer
person_id, use this:

ActiveRecord::Base.primary_key_prefix_type = :table_name_with_underscore

Plural tables make more sense to me personally.  Each row in a people
table is a Person.  It doesn''t always fit, so I take a deep breath,
write set_table_name ''person'', and move on.

We''re approaching v1.0 and there are many developers and thousands of
deployed apps.  It''s not really the time to make this kind of breaking
change because we''re running into edge cases with octopodes and viri. 
Well, on the flipside, maybe now is as good a time as there ever will
be.  We all managed the move to routes and the addition of the
DoubleRenderError.

-- 
rick
http://techno-weenie.net

Good point, I agree.

However, I''m not really convinced there''s anything broken
anyway. How
hard is it to insert the line:

ActiveRecord::Base.pluralize_table_names= false

into your environment.rb file?

On 8/10/05, Rob Park <rbpark-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> This is going about it backwards. The point of getting rid of
> pluralisation is so that we don''t have to waste processing power
> computing the plural of the word. So the code should by default check
> for the *singular* table name, and if it''s not found, *then* fall
back
> to the pluralised table name.

Thanks for the tips!  I''m not going to be too greedy about how Rails
works because I''m just glad to have it, but when the topic comes up I
started remembering my pain :)



On 8/9/05, Rick Olson
<technoweenie-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> > Why is the system necessary?  Why do table names need to be plural?
> >
> > For example, before Rails I would call my table "person",
because it
> > defines a person. Yes it hold a collection of persons, but the table
> > definition describes a person.  I''ve always been comfortable
saying
> > "select * from person where lastname like
''Thom%''.  To me it''s the
> > "person table".
> 
> I think it''s because it''s what David prefers, and it
works with most
> cases.  But, it''s not necessary by any means.  Just add this to
your
> environment.rb:
> 
> ActiveRecord::Base.pluralize_table_names = false
> 
> If you have a beef with the id primary keys and would prefer
> person_id, use this:
> 
> ActiveRecord::Base.primary_key_prefix_type = :table_name_with_underscore
> 
> Plural tables make more sense to me personally.  Each row in a people
> table is a Person.  It doesn''t always fit, so I take a deep
breath,
> write set_table_name ''person'', and move on.
> 
> We''re approaching v1.0 and there are many developers and thousands
of
> deployed apps.  It''s not really the time to make this kind of
breaking
> change because we''re running into edge cases with octopodes and
viri.
> Well, on the flipside, maybe now is as good a time as there ever will
> be.  We all managed the move to routes and the addition of the
> DoubleRenderError.
> 
> --
> rick
> http://techno-weenie.net
> _______________________________________________
> Rails mailing list
> Rails-1W37MKcQCpIf0INCOvqR/iCwEArCW2h5@public.gmane.org
> http://lists.rubyonrails.org/mailman/listinfo/rails
>

Michael Teter wrote:
>I may be asking for it, but why even bother with auto-pluralization? 
>It seems like an unnecessary gimmick, in contrast to the Rails
>philosophy of keeping things simple.  And not only does it add
>complexity that''s (imo) not necessary, it adds confusion in some
>cases.  I''m just guessing, but I imagine ESL speaking people might
>have even more problems early on getting used to it.
>
>Why is the system necessary?  Why do table names need to be plural?
>
>For example, before Rails I would call my table "person", because
it
>defines a person. Yes it hold a collection of persons, but the table
>definition describes a person.  I''ve always been comfortable saying
>"select * from person where lastname like ''Thom%''. 
To me it''s the
>"person table".
>_______________________________________________
>Rails mailing list
>Rails-1W37MKcQCpIf0INCOvqR/iCwEArCW2h5@public.gmane.org
>http://lists.rubyonrails.org/mailman/listinfo/rails
>
>  
>
Agreed.

I know almost nothing about internationalization but a little bit about 
foreign languages.  What does the pluralization do with non-English 
words?  Adding ''s'', ''es'', etc.
can''t be a very universally applicable rule.

Or do we expect everyone to Anglicize their table names?

Ben Myles wrote:
>Good point, I agree.
>
>However, I''m not really convinced there''s anything broken
anyway. How
>hard is it to insert the line:
>
>ActiveRecord::Base.pluralize_table_names= false
>
>into your environment.rb file?
>
>  
>
Not hard for most people on this list.  But non-trivial for someone 
giving Rails the once-over.

I for one want Rub/Rails to get as wide an acceptance as possible.  
Anything that trips up the first-timer/evaluator isn''t good, IMO.  How 
many reviews/blog entries have we read that have several, "yes but.." 
statements about some of the less conventional conventions.

On 8/9/05, Rick Olson
<technoweenie-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> > Why is the system necessary?  Why do table names need to be plural?
> >
> > For example, before Rails I would call my table "person",
because it
> > defines a person. Yes it hold a collection of persons, but the table
> > definition describes a person.  I''ve always been comfortable
saying
> > "select * from person where lastname like
''Thom%''.  To me it''s the
> > "person table".
> 
> I think it''s because it''s what David prefers, and it
works with most
> cases.  But, it''s not necessary by any means.  Just add this to
your
> environment.rb:
> 
> ActiveRecord::Base.pluralize_table_names = false
> 
> If you have a beef with the id primary keys and would prefer
> person_id, use this:
> 
> ActiveRecord::Base.primary_key_prefix_type = :table_name_with_underscore
> 
+1

If you don''t like it, don''t use it.  

> Plural tables make more sense to me personally.  Each row in a people
> table is a Person.  It doesn''t always fit, so I take a deep
breath,
> write set_table_name ''person'', and move on.
+1

This has worked well for people so far.  If you have an edge case,
just take the 2 seconds and type in the name explicitly....

> We''re approaching v1.0 and there are many developers and thousands
of
> deployed apps.  It''s not really the time to make this kind of
breaking
> change because we''re running into edge cases with octopodes and
viri.
> Well, on the flipside, maybe now is as good a time as there ever will
> be.  We all managed the move to routes and the addition of the
> DoubleRenderError.
I certainly think the time to make any sort of big changes like this
is PRE 1.0 release.

One would hope that people with deployed production rails apps will
not go willy nilly upgrading rails versions anyway - either by
specifying a specific gem version or by having a copy in the
lib/vendor dir.  This ensures that once it works, it will always work.

Anyone who writes web apps in Java will know this.  Any library you
use will likely be copied into your application, rather than relying
on shared copies present in the app server.  Take a look at how many
copies of log4j exist on one java server and you''ll understand :)

If the biggest problems Rails has now are little annoyances like
pluralization, things are looking pretty good.


Jason

On 8/9/05, Rob Park <rbpark-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> [...] The point of getting rid of
> pluralisation is so that we don''t have to waste processing power
> computing the plural of the word. [...]
Are you serious?

Can you provide any data, benchmark, whatever, to show that the
pluralization of names has a negative impact on the performance of
your application?  This doesn''t seem to be a very strong argument for
changing the current behavior.  But I''ll gladly rescind my comment if
proven wrong.


Jason

On 8/8/05, Kev Jackson
<kevin.jackson-1n8Jz8fJjQ+aMJb+Lgu22Q@public.gmane.org>
wrote:
> 
> >
> > It needs to be said: Auto pluralization and singularization are evil.
> > If you disagree, please look at this patch:
> >
> > <http://dev.rubyonrails.com/attachment/ticket/1844/
> > big_inflector_fix.patch>
> >
> > Did you know "schemata" is the plural of "schema"?
And that if you
> > name your class "Virus", Rails will look for the table named
"viri"?
> > The two dictionaries I consulted seem to think the plural is
> > "viruses". Stupid dictionaries!
> 
> Well I actually wrote that patch to address a few bugs that were on the
> buglist - so do you address the bugs that people post or not?  I thought
> it''d be worthwhile fixing some/most of the requests, but
it''s not up to
> me to actually include that code in the rails code-base.  This is how
> open source works (at least in all the projects I''ve participated
in),
> people contribute, and a committer decides if the contribution addresses
> a need or fixes an outstanding bug.
> 
> As for Virus/Viri.  The fact is that virus does not have a recognized
> plural form in it''s original language (I think Latin, but
I''m not
> positive).  Viruses is the ''anglicized'' version.  Rails
(at least my
> local version :) now uses ''viruses'' as the plural of
virus.  If you want
> to know more about algorithmic pluralization/singularization, then there
> are papers written on the topic and there is indeed a perl module that
> encapsulates the rules (we have at least 50% of these in the inflector
> right now, I''d actually say that in the patch we get close to 75%,
> catching the others is difficult as it depends on context).  Schema ->
> Schemata, is very very odd, but then what is the plural of Octopus?
> It''s actually Octopodes (not Octopi - I have to shout at my old
biology
> teacher about not knowing that one).  So should we use something that is
> correct in terms of the roots of the language (Octopus come from greek I
> think), or should we go with common usage?  Of course the simple answer
> is to do zero pluralization and then everything is fine (or broken, but
> all in the same way).
So, dictionary.com is wrong with their pluralization of Octopus? 
http://www.answers.com/octopus

> So, dictionary.com is wrong with their pluralization of Octopus?
> http://www.answers.com/octopus
>
Is someone writing a marine biology app?

-- 
rick
http://techno-weenie.net

> So, dictionary.com is wrong with their pluralization of Octopus? 
> http://www.answers.com/octopus
People (or persons)...

You can turn off pluralization if you don''t like it.  If you have a
more
intelligent way that rails can behave by default, submit a patch or 
write a ticket describing the change.

I would appreciate VERY MUCH if the person who is building an 
application that requires an octopuses table could email me off list and 
let me know more about what it is that they''re doing and why.

-Matt

Logically, computing plurals of words absolutely must take more time
than not doing it :)  Of course it''s probably not _much_ extra time...

But anyway, perhaps he was speaking of human computation - making sure
you understand how your words are going to be pluralized.

The fact that (I now know) there is an easy way to disable the
pluralization may make the whole discussion moot, at least for those
of us who would choose not to use it.

On 8/9/05, Jason Foreman
<threeve.org-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> On 8/9/05, Rob Park <rbpark-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> > [...] The point of getting rid of
> > pluralisation is so that we don''t have to waste processing
power
> > computing the plural of the word. [...]
> 
> Are you serious?
> 
> Can you provide any data, benchmark, whatever, to show that the
> pluralization of names has a negative impact on the performance of
> your application?  This doesn''t seem to be a very strong argument
for
> changing the current behavior.  But I''ll gladly rescind my comment
if
> proven wrong.
> 
> 
> Jason
> _______________________________________________
> Rails mailing list
> Rails-1W37MKcQCpIf0INCOvqR/iCwEArCW2h5@public.gmane.org
> http://lists.rubyonrails.org/mailman/listinfo/rails
>

Rick Olson wrote:
>>So, dictionary.com is wrong with their pluralization of Octopus?
>>http://www.answers.com/octopus
>>
>>    
>>
>
>Is someone writing a marine biology app?
>
>  
>
funny we had this exact discussion two weeks ago...
or was it three...




_______________________________________________
Rails mailing list
Rails-1W37MKcQCpIf0INCOvqR/iCwEArCW2h5@public.gmane.org
http://lists.rubyonrails.org/mailman/listinfo/rails

> 
> The goal is not to keep running the <1000 web apps that are currently
> built atop Rails, but to make Rails a compelling platform for a
> million more.
> 
Pluralisation helps greatly with this. Come out of your microcosm. 

All rails needs is a simple way to add special rules or exceptions to
the inflector tables form perhaps the environment.rb.

-- 
Tobi
http://www.snowdevil.ca - Snowboards that don''t suck
http://typo.leetsoft.com - Open source weblog engine
http://blog.leetsoft.com - Technical weblog

In article
<ea7102f90508082019155763af-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>,
tyler.kiley-
Re5JQEeQqe8AvxtiuMwx3w-XMD5yJDbdMReXY1tMh2IBg@public.gmane.org
says...
> The reasoning behind auto pluralization is still valid, and it does
> provide the benefit it claims to provide (whether you care about that
> benefit is a different issue entirely).  Why not just leave things the
> way they are right now, and just make sure newbies know how to
> manually specify the table name of a model, for those cases where the
> inflector falls short?
+1

-- 
Jay Levitt                | 
Wellesley, MA             | I feel calm.  I feel ready.  I can only
Faster: jay at jay dot fm | conclude that''s because I don''t
have a
http://www.jay.fm         | full grasp of the situation. - Mark Adler

> > The reasoning behind auto pluralization is still valid, and it does
> > provide the benefit it claims to provide (whether you care about that
> > benefit is a different issue entirely).  Why not just leave things the
> > way they are right now, and just make sure newbies know how to
> > manually specify the table name of a model, for those cases where the
> > inflector falls short?
> 
> +1
+20

-- 
Cheers

Koz

+20 as well :-)

On 8/9/05, Michael Koziarski
<koziarski-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> > > The reasoning behind auto pluralization is still valid, and it
does
> > > provide the benefit it claims to provide (whether you care about
that
> > > benefit is a different issue entirely).  Why not just leave
things the
> > > way they are right now, and just make sure newbies know how to
> > > manually specify the table name of a model, for those cases where
the
> > > inflector falls short?
> >
> > +1
> +20
> 
> --
> Cheers
> 
> Koz
> _______________________________________________
> Rails mailing list
> Rails-1W37MKcQCpIf0INCOvqR/iCwEArCW2h5@public.gmane.org
> http://lists.rubyonrails.org/mailman/listinfo/rails
> 

-- 
- Aaron ''Jomdom'' Ransley
- Web: www.jomdom.net
- Mail: jomdom-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org

Merriam Webster says the plural of Octopus is "octopuses or octopi"

www.m-w.com

Carl

On 8/8/05, Kev Jackson
<kevin.jackson-1n8Jz8fJjQ+aMJb+Lgu22Q@public.gmane.org> wrote:
context).  Schema ->
> Schemata, is very very odd, but then what is the plural of Octopus?
> It''s actually Octopodes (not Octopi - I have to shout at my old
biology
> teacher about not knowing that one).  So should we

I agree.

On 8/9/05, Aaron Ransley <jomdom-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> +20 as well :-)
> 
> On 8/9/05, Michael Koziarski
<koziarski-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> > > > The reasoning behind auto pluralization is still valid, and
it does
> > > > provide the benefit it claims to provide (whether you care
about that
> > > > benefit is a different issue entirely).  Why not just leave
things the
> > > > way they are right now, and just make sure newbies know how
to

On Aug 9, 2005, at 1:37 PM, Tobias Luetke wrote:
>>
>> The goal is not to keep running the <1000 web apps that are
currently
>> built atop Rails, but to make Rails a compelling platform for a
>> million more.
>>
>>
>
> Pluralisation helps greatly with this. Come out of your microcosm.
You''re encouraging people to learn the wrong way to name tables. We  
have a duty to NOT perpetuate ignorance and incompetence.

Regards,
Ed

-- 
Transmogrify, LLC * <http://xmog.com/>

> > Pluralisation helps greatly with this. Come out of your microcosm.
> 
> You''re encouraging people to learn the wrong way to name tables.
We
> have a duty to NOT perpetuate ignorance and incompetence.
Please get some perspective. 

There''s no right or wrong way to name tables,  just different models
with different options with pros and cons.  Choosing to name tables
differently doesn''t make someone ignorant or incompetent.



-- 
Cheers

Koz

On 8/9/05, Rick Olson
<technoweenie-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> > So, dictionary.com is wrong with their pluralization of Octopus?
> > http://www.answers.com/octopus
> >
> 
> Is someone writing a marine biology app?
+1 zinger

>> Pluralisation helps greatly with this. Come out of your microcosm.
> 
> You''re encouraging people to learn the wrong way to name tables.
We
> have a duty to NOT perpetuate ignorance and incompetence.
Let''s play nice here. There''s no "wrong" way. Lots
of people were raised
to make them singular (myself included), lots of people to make them plural.

There continues to be active debate on which is more "natural", and 
incompetence plays no larger role than it does in the vi vs. emacs wars.

It''s also incredibly easy to make it work the way you want, just by
setting:

	ActiveRecord::Base.pluralize_table_names = false

Given that using plural table names is a common convention (even if
it''s
not the one I happen to use), I consider auto-pluralization to be a 
CRITICAL feature of Rails -- even though the first thing I do is turn it 
off! Rails is all about convention over configuration, and it''s a 
feature that makes it work for lots of folks. My convention just
doesn''t
happen to require any special feature, other than to turn this one off.

Now you COULD argue that singular table names is more common, and that 
we should change the default. That would break existing apps, though the 
"fix" would be as simple as:

	ActiveRecord::Base.pluralize_table_names = true

But when it comes down to it, does it really matter? Probably the most 
important thing is that the key assumptions on conventions (table names, 
primary key names, etc) are clearly identified along with the associated 
knobs, so that new folks can make it work for them very easily.

Perhaps the generated environment.rb file should include a few of these 
knobs with some useful comments.

> You''re encouraging people to learn the wrong way to name tables.
We
> have a duty to NOT perpetuate ignorance and incompetence.
Please use struts.

-- 
Tobi
http://www.snowdevil.ca - Snowboards that don''t suck
http://typo.leetsoft.com - Open source weblog engine
http://blog.leetsoft.com - Technical weblog

> Let''s play nice here. There''s no "wrong" way.
Lots of people were
> raised to make them singular (myself included), lots of people to make 
> them plural.
>
> There continues to be active debate on which is more "natural",
and
> incompetence plays no larger role than it does in the vi vs. emacs wars.
>
> It''s also incredibly easy to make it work the way you want, just
by
> setting:
>
>     ActiveRecord::Base.pluralize_table_names = false
>
> Given that using plural table names is a common convention (even if 
> it''s not the one I happen to use), I consider auto-pluralization
to be
> a CRITICAL feature of Rails -- even though the first thing I do is 
> turn it off! Rails is all about convention over configuration, and 
> it''s a feature that makes it work for lots of folks. My convention
> just doesn''t happen to require any special feature, other than to
turn
> this one off.
>
> Now you COULD argue that singular table names is more common, and that 
> we should change the default. That would break existing apps, though 
> the "fix" would be as simple as:
>
>     ActiveRecord::Base.pluralize_table_names = true
>
> But when it comes down to it, does it really matter? Probably the most 
> important thing is that the key assumptions on conventions (table 
> names, primary key names, etc) are clearly identified along with the 
> associated knobs, so that new folks can make it work for them very 
> easily.
>
> Perhaps the generated environment.rb file should include a few of 
> these knobs with some useful comments.
>
>
The way SOME tools do this is that they have the config file 
(environment.rb) pre-populated with the defaults, to make it easy to 
know how to switch them without having to look up the names.   This may 
be the best solution.   Even when you turn that off, it still does funny 
things with model and controller names.    Is there a way to switch off 
all pluralization?   I had a test table called "bogus" it make lots of
"bogu"  stuff!

Any way, this is funny, and I would prefer to have it just take what I 
type, but at the same time I understand no breaking existing apps.   
Just pre-populate environment.rb with 20 or 30 of the interesting or 
popular  items so they are easy to change, and everyone is happy, even 
the octopuseseseseses.

Caleb Buxton wrote:
>On 8/9/05, Rick Olson
<technoweenie-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>  
>
>>>So, dictionary.com is wrong with their pluralization of Octopus?
>>>http://www.answers.com/octopus
>>>
>>>      
>>>
>>Is someone writing a marine biology app?
>>    
>>
nope,

http://search.cpan.org/src/DCONWAY/Lingua-EN-Inflect-1.89/lib/Lingua/EN/Inflect.pm

see:

 >>

I''m not taking any further correspondence on:

=over

=item "octopi".

Despite the populist pandering of certain New World dictionaries, the
plural is "octopuses" or (for the pendantic classicist)
"octopodes". The
suffix "-pus" is Greek, not Latin, so the plural is
"-podes", not "pi".


=item "virus".

Had no plural in Latin (possibly because it was a mass noun).
The only plural is the Anglicized "viruses".

<<

If I''d have known that people would have taken exception to an attempt 
to fix reported bugs I wouldn''t have bothered. Talk about biting the 
hand that feeds.
And as it''s a one-liner to turn this feature off I think this whole 
thread has been a waste of bits, especially as we now have people 
claiming that using plural table names is ignorant, as opposed to simply 
a different style.

I''ll think very carefully before contributing a patch to Rails again if
this is the sort of vehement opposition it brings out of the woodwork - 
and as far as I''m aware the damn patch hasn''t even been
committed, so
it''s irrelevant.

Kev

On Aug 9, 2005, at 7:11 PM, Michael Koziarski wrote:
>>> Pluralisation helps greatly with this. Come out of your microcosm.
>>>
>> You''re encouraging people to learn the wrong way to name
tables. We
>> have a duty to NOT perpetuate ignorance and incompetence.
>>
>
> Please get some perspective.
>
> There''s no right or wrong way to name tables,  just different
models
> with different options with pros and cons.  Choosing to name tables
> differently doesn''t make someone ignorant or incompetent.
There may not be one right way, but there are most definitely wrong  
ways. This isn''t high school English class where the teacher is  
obligated to find something positive to say about everyone''s  
interpretation of Antigone.


On Aug 9, 2005, at 7:34 PM, Michael Schoen wrote:
>>> Pluralisation helps greatly with this. Come out of your microcosm.
>>>
>> You''re encouraging people to learn the wrong way to name
tables.
>> We  have a duty to NOT perpetuate ignorance and incompetence.
>>
>
> Let''s play nice here. There''s no "wrong" way.
Lots of people were
> raised to make them singular (myself included), lots of people to  
> make them plural.
You were raised by people who taught you well.
> There continues to be active debate on which is more "natural",
and
> incompetence plays no larger role than it does in the vi vs. emacs  
> wars.
This isn''t a question of what feels natural: It''s a question
of not
indulging their bad habits or misconceptions.


On Aug 9, 2005, at 8:20 PM, Tobias Luetke wrote:
>>> Pluralisation helps greatly with this. Come out of your microcosm.
>>>
>> You''re encouraging people to learn the wrong way to name
tables. We
>> have a duty to NOT perpetuate ignorance and incompetence.
>>
>
> Please use struts.
I''ve been insulted more coherently by more knowledgeable people than  
you. Struts knows nothing about how a model object is stored in a  
database -- if it''s stored in a database at all. I''ve never
written a
Struts application. I''m generally program in Scheme, Python, Perl,  
Objective-C, and Ruby: five dynamic languages. When I learn how to do  
something, I try to figure out how smart, experienced people do it.  
I''ve never met a smart, experienced database developer who would name  
a table "posts" even if someone put a gun to his or her head.

This is a pointless battle -- and why do I care? Go ahead and name  
your tables "people", "viri", and "octopodes".
I''ll just continue
adding my magic configuration lines in every app I build.

Regards,
Ed

-- 
Transmogrify, LLC * <http://xmog.com/>

> This is a pointless battle -- and why do I care? Go ahead and name
> your tables "people", "viri", and
"octopodes". I''ll just continue
> adding my magic configuration lines in every app I build.
Yes why do you care?   Stop flaming.  

Can we please just put this thread to death.

-- 
Cheers

Koz

On 8/9/05, Michael Teter
<michael.teter-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> Logically, computing plurals of words absolutely must take more time
> than not doing it :)  Of course it''s probably not _much_ extra
time...
Took the words right out of my mouth, thank you.

My only complaint with auto pluralising words is that it gets a lot of
words wrong (in fact there are many words which we can never agree on
the correct pluralisation, so any automatic pluralisation routine can
never been 100% correct because nobody really knows what 100% correct
is).

As far as I can see, the only "benefit" of having auto pluralisation
is so that some rails code reads more like english than code, which is
a dubious benefit at best. Code is code, it should be precise, not
like English which is vague and imprecise.

Of course, having pluralisation turned off by default in future
releases will only break all the existing applications, which is a
cure worse than the disease, so for now I''ll just have to consider
auto-pluralisation to be a crufty feature necessary for compatibility
that should be deprecated sooner or later and phased out.

-- 
Urban Artography
http://artography.ath.cx

Rob Park <rbpark-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> writes:

> Of course, having pluralisation turned off by default in future
> releases will only break all the existing applications, which is a
> cure worse than the disease,...
I disagree.  There''s no better time than RIGHT NOW than to fix this
arbitrary and easily most confusing aspect of rails.  There are never
going to be /fewer/ rails apps out than there are now, and with the
PragBook just having hit the streets, I suspect a big spike in numbers
coming soon.


-- 
I tend to view "truly flexible" by another term: "Make everything
equally hard". -- DHH

On 14 Aug 2005 17:44:53 -0400, Michael Campbell
<michael.campbell-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> I disagree.  There''s no better time than RIGHT NOW than to fix
this
> arbitrary and easily most confusing aspect of rails.  There are never
> going to be /fewer/ rails apps out than there are now, and with the
> PragBook just having hit the streets, I suspect a big spike in numbers
> coming soon.
The original reasoning behind this feature (more conversational
discussion of app development) is still valid.  If you''ve read the
pragbook, you know that it devotes ample time to explaning how to
disable the feature, and how to work around it when it pluralizes
incorrectly.  If you don''t like it, you can disable it; what''s
your
beef with that?

Think about your reasoning for wanting to make the change now -- the
prag book just started shipping, so thousands of rails devs are
getting dead-tree documentation for a feature that you want to change
after the ink has barely had time to dry.  Does that really make
sense?

If there was ever a time to "fix" this "problem", it was
three or four
months ago; now, the best option is to just get in the habit of
disabling it in your apps if it''s that much of an inconvenience to
you.

(Disclaimer:  I happen to like auto pluralization.)

Tyler

I haven''t looked at the code base with this in mind, but I''m
curious
- how much code is loaded and parsed even if pluralization is  
disabled? How much memory is used? Has anyone benchmarked application  
start up and per-request stats for pluralization being enabled,  
disabled, and removed?




On Aug 14, 2005, at 8:12 PM, Tyler Kiley wrote:
> On 14 Aug 2005 17:44:53 -0400, Michael Campbell
> <michael.campbell-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>
>> I disagree.  There''s no better time than RIGHT NOW than to fix
this
>> arbitrary and easily most confusing aspect of rails.  There are never
>> going to be /fewer/ rails apps out than there are now, and with the
>> PragBook just having hit the streets, I suspect a big spike in  
>> numbers
>> coming soon.
>>
>
> The original reasoning behind this feature (more conversational
> discussion of app development) is still valid.  If you''ve read the
> pragbook, you know that it devotes ample time to explaning how to
> disable the feature, and how to work around it when it pluralizes
> incorrectly.  If you don''t like it, you can disable it;
what''s your
> beef with that?
>
> Think about your reasoning for wanting to make the change now -- the
> prag book just started shipping, so thousands of rails devs are
> getting dead-tree documentation for a feature that you want to change
> after the ink has barely had time to dry.  Does that really make
> sense?
>
> If there was ever a time to "fix" this "problem", it
was three or four
> months ago; now, the best option is to just get in the habit of
> disabling it in your apps if it''s that much of an inconvenience to
> you.
>
> (Disclaimer:  I happen to like auto pluralization.)
>
> Tyler
> _______________________________________________
> Rails mailing list
> Rails-1W37MKcQCpIf0INCOvqR/iCwEArCW2h5@public.gmane.org
> http://lists.rubyonrails.org/mailman/listinfo/rails
>

Let me weigh in here. I just ordered the "dead-tree" document, as 
someone just put it. I''ve been a programmer for 43 years. In that 43 
years I''ve seen languages come and go, computer architectures come and 
go, operating systems come and go, and applications come and go. I''ve 
decided to learn Ruby, Ruby on Rails and web application development 
simultaneously/concurrently.

/*Let me emphasize this -- I want this heard loud and clear -- a 
development environment exists for the sole purpose of making the 
developer''s job of satisfying the end users as easy as possible. */A 
development environment where someone can make an arbitrary change in 
the syntax or semantics without good reason makes that hard. I wrote 
some Perl 4 code a few years ago, which ran perfectly well until someone 
changed Perl to use character semantics rather than byte semantics by 
default. And I won''t trouble you with the troubles gcc 4.0 hath
wrought;
I''m still at 3.3.5 for a very good reason.

Ruby is a young language -- 1.8 vs. Perl 5.8. And Ruby on Rails is a 
young framework. Ruby and Ruby on Rails owe it to the developers and 
their customers to maintain a core environment that will work 
/***forever***/, and clearly document stuff that isn''t part of that 
core. Then it''s *my* responsibility as a developer to use only what I 
know is going to be there for me.

Tyler Kiley wrote:
>The original reasoning behind this feature (more conversational
>discussion of app development) is still valid.  If you''ve read the
>pragbook, you know that it devotes ample time to explaning how to
>disable the feature, and how to work around it when it pluralizes
>incorrectly.  If you don''t like it, you can disable it;
what''s your
>beef with that?
>
>Think about your reasoning for wanting to make the change now -- the
>prag book just started shipping, so thousands of rails devs are
>getting dead-tree documentation for a feature that you want to change
>after the ink has barely had time to dry.  Does that really make
>sense?
>
>If there was ever a time to "fix" this "problem", it was
three or four
>months ago; now, the best option is to just get in the habit of
>disabling it in your apps if it''s that much of an inconvenience to
>you.
>

_______________________________________________
Rails mailing list
Rails-1W37MKcQCpIf0INCOvqR/iCwEArCW2h5@public.gmane.org
http://lists.rubyonrails.org/mailman/listinfo/rails

On Aug 14, 2005, at 8:12 PM, Tyler Kiley wrote:
> On 14 Aug 2005 17:44:53 -0400, Michael Campbell
> <michael.campbell-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>
>> I disagree.  There''s no better time than RIGHT NOW than to fix
this
>> arbitrary and easily most confusing aspect of rails.  There are never
>> going to be /fewer/ rails apps out than there are now, and with the
>> PragBook just having hit the streets, I suspect a big spike in  
>> numbers
>> coming soon.
>>
>
> The original reasoning behind this feature (more conversational
> discussion of app development) is still valid.
I would be much happier with the status quo if environment.rb had a  
couple lines setting the key- and table-naming conventions to the  
stock values with a comment alluding to other possible values. This  
issue bothers me because it took so much effort to find out how to  
change the settings.

Also, with singular table names and PKs prefixed with the table name,  
the stock database session support doesn''t work. I had to create my  
own session model.

Regards,
Ed

-- 
Transmogrify, LLC * <http://xmog.com/>

-10.  I (and I believe the silent majority) have found auto-plural a 
joyful ''nice touch''.

If you disagree then please put up a wiki page linked to the one below 
describing your philosophy on why you think pluralization should be off 
by default and what the alternate naming convention should be.  Then 
others of like-mind will have a resource.

Unless you''re submitting a patch, the debate here is pointless and just
going to be buried in the archives in a month.

http://wiki.rubyonrails.com/rails/show/HowtoDisableAutoTableNamePluralisation


Michael Campbell wrote:
> Rob Park <rbpark-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> writes:
> 
> 
> 
>>Of course, having pluralisation turned off by default in future
>>releases will only break all the existing applications, which is a
>>cure worse than the disease,...
> 
> 
> I disagree.  There''s no better time than RIGHT NOW than to fix
this
> arbitrary and easily most confusing aspect of rails.  There are never
> going to be /fewer/ rails apps out than there are now, and with the
> PragBook just having hit the streets, I suspect a big spike in numbers
> coming soon.
> 
>

@rob park
Java!  if you want strict, to the point, and *more* clearity in your
software go use java!  autocomplete is a feature, and as tyler said
can just as easily be disabled.

as for how much time it adds to pluralize words id imagine none at
all.  in production mode id *hope* that the pluralized phrases are
computed once.  so you wait .001 seconds longer for your application
to start it, i think we can wait.

personally i believe auto-pluralize is a nice feature and for most new
people even better for them.  your attempting to turn ''convention over
configuration'' into ''conventionS'' and we dont want
that.  i dont want
to look at another rails app and go "oh wth, oh they have this off, oh
wait what, oh they turned this on, wait shit, and the changed this"
running through code chewing your head off looking for why shit dont
work is even more worse then wasting your time scrolling through a
configuration file.


@ed watkeys
i think i would submit this in as a bug, instead of just sitting on it
assuming rails is broke and no one will ever fix it :-/

the only way something will get better/fixed is if someone complains about it.


On 8/15/05, San <sanzbox-/E1597aS9LQAvxtiuMwx3w@public.gmane.org>
wrote:
> -10.  I (and I believe the silent majority) have found auto-plural a
> joyful ''nice touch''.
> 
> If you disagree then please put up a wiki page linked to the one below
> describing your philosophy on why you think pluralization should be off
> by default and what the alternate naming convention should be.  Then
> others of like-mind will have a resource.
> 
> Unless you''re submitting a patch, the debate here is pointless and
just
> going to be buried in the archives in a month.
> 
>
http://wiki.rubyonrails.com/rails/show/HowtoDisableAutoTableNamePluralisation
> 
> 
> Michael Campbell wrote:
> > Rob Park <rbpark-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
writes:
> >
> >
> >
> >>Of course, having pluralisation turned off by default in future
> >>releases will only break all the existing applications, which is a
> >>cure worse than the disease,...
> >
> >
> > I disagree.  There''s no better time than RIGHT NOW than to
fix this
> > arbitrary and easily most confusing aspect of rails.  There are never
> > going to be /fewer/ rails apps out than there are now, and with the
> > PragBook just having hit the streets, I suspect a big spike in numbers
> > coming soon.
> >
> >
> 
> _______________________________________________
> Rails mailing list
> Rails-1W37MKcQCpIf0INCOvqR/iCwEArCW2h5@public.gmane.org
> http://lists.rubyonrails.org/mailman/listinfo/rails
> 

-- 
Zachery Hostens <zacheryph-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>

I know the horse is pretty dead, but...

Everyone should be able to agree that the auto-pluralizer will never
be 100% correct in the decisions it makes.  Would anyone disupte that?
 It follows that the pluralization system CAN be a stumbling point - a
point of confusion for developers.

I fully understand the arguments (dead tree book exists being the most
significant) for why we should keep the current system, but that''s
about the only logical reason.

Plain and simple, if there were no pluralization system, there would
be fewer surprises.  I think it was the Python people who pushed the
concept of least surprise (is best), and it was a literature professor
of mine who pushed the concept of removing words/phrases (features in
this case) that are stumbling blocks for the reader (developer, code
reviewer). Does the feature add enough convenience or benefit to
outweigh the dangers/confusion?

It''s difficult to draw a hard line between practical helper features
and frivolous features, so I can understand how this system made it
this far.  But at this point I believe it should be fairly clear that
the pluralization system, if it should exist at all, should be OFF by
Default.  Yes it would break existing sites, but Rails is new enough
that chances are good the developers of those sites are still paying
attention to What''s New In Rails, and thus they''d understand
the need
to add the one line of config to Enable the pluralization feature.

If we need some reason to be fearful of dangerous or difficult-to-use
features, just look at the history of C++, a language where you should
buy and read at least two books just to learn what NOT to do with the
language (Scott Meyers - Effective C++ series).

On 8/15/05, Michael Teter
<michael.teter-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> I know the horse is pretty dead, but...
+1E100

(that''s scientific notation, not hex)

Tyler

On 8/15/05, Zachery Hostens
<zacheryph-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:
> @rob park
> Java!  if you want strict, to the point, and *more* clearity in your
> software go use java!  autocomplete is a feature, and as tyler said
> can just as easily be disabled.
Programmer programmer
Programmer.new(User.new(Student.new(Person.new(String.new("Rob
Park"))))).addHatedObject((HatedObject)String.new("Java"));
programmer.hate();

When I was a CompSci student at university, Java was the language used
to teach the first year''s courses and it was a really annoying and
tedious language to try and do anything at all with.

What I think should be done is change rails so that if pluralisation
isn''t set in environment.rb then it defaults to on, but new versions
of environment.rb should come with a line disabling it. That way old
apps being upgraded will still have pluralisation by default but new
apps will have it turned off. Then when 2.0 comes out it can be set to
disabled by default.

-- 
Urban Artography
http://artography.ath.cx