Rails - Dec 2005 - shared hostings: how to compare/evaluate them (aka: "Rails needs a index")

Hi all,


New offers for cheap Rails hosting keep popping-up at an increasing pace.

    - Are they any good?
    - Is this 60$/year plan good enough for me?
    - "Is DreamHost any good, despite the hundreds of accounts per
server?"
    - Will plan Z support the load/traffic my client expect?
    - Is this 60$/month plan @ hoster X better than the 14$/month plan @ 
hoster Y
    - Why is my app 3 times slower than last months? Is it my code, or 
is it the server?
    - etc..
All those questions are vital when comes deployment time and today, when 
hosting price really matters, we''re clueless.

AFAIK, the only way to an answer those questions is for through a 
standard (set of) Rails application(s) that would exercise and measure 
various areas of Rails, and publish the results interactively or send 
them to a repository where they would be accumulated. We would just have 
to install one instance at each hoster (or better, each server).

If the community''s pressure is sufficient, we could even have the 
hosters do this on a voluntary basis. It would benefit the good ones, as 
they could publish their ''Rails Index'' on the plans ad pages.


Rails is right on many points where mature platforms are still struggling
   - deployment
   - testing
   - plugins
   - upgrades
   - simplicity

this is one more area where it can "Think Different".


What say you?

Alain

>AFAIK, the only way to an answer those questions is for through a
>standard (set of) Rails application(s) that would exercise and measure
>various areas of Rails, and publish the results interactively or send
>them to a repository where they would be accumulated. We would just have
>to install one instance at each hoster (or better, each server).
>If the community''s pressure is sufficient, we could even have the
>hosters do this on a voluntary basis. It would benefit the good ones, as
>they could publish their ''Rails Index'' on the plans ad
pages.
> What say you?
I say "Why are there no responses to this very good 6-month old idea?"

It seems like there should be a standard test module that users could
decide whether or not to run or make accessible.  Then for every Rails
site that had it running, you could just go to railssite.com/benchmark
or something and glean the data.  Someone could write a nice Rails app
to ping sites known to be running the benchmark, and graphs could be
made to codify the data.

Am I tripping here or is this not a good idea?

I am looking for a host right now, and having no prior experience with
hosts that have Rails I''m stuck.  The RoR wiki lists hosts; only one
has
a user experience associated with it.  Pick any name on that list and
google "railshostname sucks" and you''re guaranteed some hits,
but who
knows if those people knew what they were doing or if their complaints
are valid in the context of my usage patterns?

-- 
Posted via http://www.ruby-forum.com/.

>>>>> "foobario" == foobario 
<foobario@yahoo.com> writes:
> It seems like there should be [...]
> Someone could write [...]
> [...] graphs could be made [...]
> Am I tripping here or is this not a good idea?
It does not matter how good or bad an idea is as long as there is
nobody willing to create, write and make the stuff needed to make it
real.

You obviously think that this is a good idea. You also feel a need for
it. The question then becomes what you''re willing to do in order to
make it happen.

The strategy you have just tried, posting about it and hoping someone
else will do it, is extremely common and has the advantage of
requiring almost no effort. The drawback is that its success rate is
vanishingly small. Not quite nonexistent, but pretty close. I''ve used
this one myself, on occasion.

Another common strategy is to pay someone to make it happen. *Much*
better success rate, but with the drawback of requiring significant
amounts of a resource most of us are short of (money, that is). This
one is mostly used by companies, as far as I can tell. 

The strategy most of us try to avoid (more or less consciously) is of
course to build it yourself. For most ideas, this requires significant
effort. Sometimes part of that effort will be learning the skills
needed to build whatever it is. But sometimes all that''s required is a
little initiative and enthusiasm, to get something started up that
others can contribute to. 
-- 
		     Calle Dybedahl <calle@cyberpomo.com>
		 http://www.livejournal.com/users/cdybedahl/
    "Ah, optimism. I remember optimism." -- Anya, Buffy the Vampire
Slayer

Foobario

    > Am I tripping here or is this not a good idea?

This is a brilliant idea :)


    > I say "Why are there no responses to this very good 6-month old
idea?"

My 2 cents: there is way too much noise and too much traffic in this forum.
My other 2 cents: core and advanced railers don''t use shared hosting,
so
they don''t care much. Look at scaffold, globalization, etc...
Rails is not for hobbyist, it is for business, serious business.

Fair enough: Rails is free.
Painful though, when those who need the info are 1/the majority, and 
2/the ones who don''t have the resources/technical knowledge to obtain 
the info.


I''m puzzled though that this very question never caught the eye of at 
least one of the core members, or of the old/experienced railers.
Rails is so much smarter when it come to testing, deploying, RJS, 
caching, templating, etc. and then there is this hole in its foot.

Alain


>> AFAIK, the only way to an answer those questions is for through a
>> standard (set of) Rails application(s) that would exercise and measure
>> various areas of Rails, and publish the results interactively or send
>> them to a repository where they would be accumulated. We would just
have
>> to install one instance at each hoster (or better, each server).
> 
>> If the community''s pressure is sufficient, we could even have
the
>> hosters do this on a voluntary basis. It would benefit the good ones,
as
>> they could publish their ''Rails Index'' on the plans
ad pages.
> 
>> What say you?
> 
> I say "Why are there no responses to this very good 6-month old
idea?"
> 
> It seems like there should be a standard test module that users could
> decide whether or not to run or make accessible.  Then for every Rails
> site that had it running, you could just go to railssite.com/benchmark
> or something and glean the data.  Someone could write a nice Rails app
> to ping sites known to be running the benchmark, and graphs could be
> made to codify the data.
> 
> Am I tripping here or is this not a good idea?
> 
> I am looking for a host right now, and having no prior experience with
> hosts that have Rails I''m stuck.  The RoR wiki lists hosts; only
one has
> a user experience associated with it.  Pick any name on that list and
> google "railshostname sucks" and you''re guaranteed some
hits, but who
> knows if those people knew what they were doing or if their complaints
> are valid in the context of my usage patterns?
>