R devel - Sep 1997 - R-alpha: Re: R-Mailing lists ok again. ..sorry..

Martin Maechler <maechler@stat.math.ethz.ch> writes:
> This caused 3 postings (1 to R-help, 2 to R-devel) to be aborted.
> For some reason, it also seems majordomo did not save the postings in
> a 'dead.letter' or something.
> So could you please re-submit them? 
> These are the sendmail entries:
> Sep 17 21:14:21 .. sendmail[9364]: VAA09363:
from=<bates@stat.wisc.edu>, size=390, class=0, pri=30390, nrcpts=1,
msgid=<6ren6nbvnt.fsf@franz.stat.wisc.edu>, proto=ESMTP,
relay=root@franz.stat.wisc.edu [128.105.5.95]
> Sep 17 21:47:26 .. sendmail[9518]: VAA09517:
from=<bates@stat.wisc.edu>, size=331, class=0, pri=30331, nrcpts=1,
msgid=<6rd8m7bu42.fsf@franz.stat.wisc.edu>, proto=ESMTP,
relay=root@franz.stat.wisc.edu [128.105.5.95]
I can't find copies of those messages so I will summarize here.  One
of them was unimportant - simply an observation that the mis-typing of
TITLE as TITLe in the documentation for the library function is
intentional.  The second was a query to Fritz with a copy to this list
on whether the data files in the bootstrap library are indeed
accessible with the data function.  It seems not.  If there was a
convention that files ending with .data in a data directory should be
read with read.table I think that would be great.  Actually I think it
would be "insanely great" if they were read with
 read.table(filename, header = TRUE)
-- 
Douglas Bates                            bates@stat.wisc.edu
Statistics Department                    608/262-2598
University of Wisconsin - Madison        http://www.stat.wisc.edu/~bates/
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To:
r-devel-request@stat.math.ethz.ch
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

>>>>> "Doug" == Douglas Bates
<bates@stat.wisc.edu> writes:
    Doug>  .....
    Doug> I can't find copies of those messages so I will summarize here.

sorry, once more.

    Doug> One of them was unimportant - simply an observation that the
    Doug> mis-typing of TITLE as TITLe in the documentation for the library
    Doug> function is intentional.  

yes! (it partly breaks the HTML version..)

But with the new Rd (man) style and its perl scripts, the problem is gone
anyway.

    Doug> The second was a query to Fritz with a
    Doug> copy to this list on whether the data files in the bootstrap
    Doug> library are indeed accessible with the data function.  It seems
    Doug> not.
This also has quite a bit changed in the current source.

As Kurt says, all the packages have to be updated though before they work
with the current directory structure.

    Doug> f there was a convention that files ending with .data in a
						          ~~~~~
    Doug> data directory should be read with read.table I think that would
    Doug> be great.
indeed!
Maybe we should have  '.dat'
in order  for the Win3.x  version to need less adaption for porting... 

    Doug> Actually I think it would be "insanely great" if they
    Doug> were read with read.table(filename, header = TRUE) 
			 -------------------  ##############
Yes again!

Martin
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To:
r-devel-request@stat.math.ethz.ch
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

>> If you're going to get into reading databases let me know. I've
already
>> done a fair amount of work on this and have a working interface to
Fame
>> time series databases (and any other database which can be made to
>> support a generic protocol I've been working on).
>I probably won't have any time for this is the near future, but perhaps
>someone else will fill in ...
>Anyway, if there is more info about the `generic protocol' you
mentioned
>that you can easily make available, the please do so.
Kurt

There is a fair amount of information and example code available at
http://www.bank-banque-canada.ca/pgilbert
Follow the link to TSPADI.

There is a working server front end for the Fame time series database
software as well as a very simplistic file based "database" we made
for
test purposes. There is also a client side interface for Splus and for a
"generic client" which is intended as an example and also for test
purposes. The version of the code on the web page needs a few minor
changes for R. I have done that and will soon make it available.

Please note that this is not a database protocol like SQL which allows
queries. The client must provide an identifier for an object and the
server simply returns the object.

The protocol uses RPC and has only been tested on TCP/IP  networks,
although I've been told RPC is available with some other networks. It
does work across the Internet. (It can also be run locally without a
network using loopback.)

An interesting project would be to write a wrapper (eg. in Perl) for
some of the data being provided as text formatted pages. Then a server
could be setup to provide the data with the PADI protocol so that it
could be used directly by application programs like R.  If anyone has
some bright student looking for something to do I can provide more
details.

Paul Gilbert

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To:
r-devel-request@stat.math.ethz.ch
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-