[prev in list] [next in list] [prev in thread] [next in thread] 

List:       gtm-perl
Subject:    Re: [gtm-perl] Re: etc :-)
From:       Jim Self <jaself () ucdavis ! edu>
Date:       2001-01-18 13:06:23
[Download RAW message or body]

Thomas Good wrote:
>On Tue, 16 Jan 2001, Jim Self wrote:
>
>> Thomas wrote:
>> >How is the Saint?  Did he get GT.M working on FBSD?  I tried and failed.
>>
>> Yes he did. It is on a computer in my office waiting for me or Ed to
test.
>
>Is he willing to pass along the recipe?  I tried to get it going under
>emulation (fbsd linux compat) but it died on me after running once...

I am sure he is. However he ported FT05 so it will need to be done again for
newer versions. Assuming that the source will be released to GPL any time
now, I suppose we should wait on that.

>
>> >  Except add perl...can't forget old Larry W!
>>
>> Perl is the first thing included in "etc" extending the work started with
>> MontyPerl. Most recently, Steve has been using Perl and the MUMPS style
>> globals module (using BerkeleyDB) as foundation for a Student/Faculty
>> evaluations application (VOLES). He says it sped up the application
>> significantly compared to his previous version which used a plain db
>> interface. I believe that replacing BerkelyDB with an interface to GTM
will
>> speed it up again and make it simpler at the same time. It will also
allow
>> our M programmers to access VOLES data from MUMPS and will provide VOLES
>> with direct access to other parts of the system.
>
>Jim - what are the big diffs between the mumps method of storing data
>and the std Oracle b*tree approach?  The more I read the less clear this
>is to me.

I am not at all familiar with Oracle, but implementing MUMPS globals in
MontyPerl on top of the BerkeleyDB B+tree access method was conceptually
very simple since that provides ordered hashes to Perl. We devised a
representation of a MUMPS global reference as a string that would sort
according to the MUMPS collation rules (canonic numbers per subscript
collate prior to all other strings) and immediately obtained the
functionality of $QUERY from the BerkeleyDB cursor positioning functions,
DB_FIRST, DB_LAST, DB_NEXT, and DB_PREV. It might be easiest for you to
understand by simply reading Steve's Perl code for the global module in
MontyPerl.

I might be able to give some comparison to Oracle's b*trees if you list some
of the features of interest.

---------------------------------------
Jim Self
Manager and Chief Developer
VMTH Computer Services, UC Davis
(http://www.vmth.ucdavis.edu/us/jaself)

[prev in list] [next in list] [prev in thread] [next in thread]