[Date Index] [Thread Index] [Author Index]

Re: General--Mathematica and Subversion

• To: mathgroup at smc.vnet.net
• Subject: [mg69668] Re: [mg69655] General--Mathematica and Subversion
• From: jmt <jmt at dxdydz.net>
• Date: Wed, 20 Sep 2006 02:44:28 -0400 (EDT)
• Organization: dxdydz
• References: <200609190945.FAA28625@smc.vnet.net>

On Tuesday 19 September 2006 11:45, you wrote:
> Hi!
>
> I want to use Subversion (http://subversion.tigris.org) to organize my
> Mathematica programs.
>
> I experienced, that the structure saved in the Mathematica notebooks
> (Cell[...], etc) does not work well with a version control system.
>
> Is there an alternative to the standard Mathematica .nb file?
>
>
> TIA
> michael

Using a version control tool for notebooks is not so difficult : the real 
trouble comes from the notebook cache and eventually the graphics.

When CVS was the only available tool for version control, a solution had been 
developped by Tim Wichmann :
http://www.itwm.fhg.de/as/asemployees/wichmann/nbcache.html
His perl script stripped out irrelevant data, and I used it with version 5.0 
without noticing errors - but that does not mean the whole process is error 
free !

When it comes to subversion, using subversion does not lead to any trouble at 
all : subversion does not have CVS limitations like needing files with end of 
line characters, and stores any file as a stream of bytes (and supposes you 
have a reasonable modern hard drive, with reasonable free space for a 
multimedia era). The difficulty arises when you want to perform 'diff' 
because then the cache is involved.

The Wolfram team is now developping a dedicated tool :

http://www.wolfram.com/products/workbench/

which provides version control. If I had no versioning tool already running 
(subversion) I for sure would give it a try.

jmt

• Follow-Ups:
  • Re: Re: General--Mathematica and Subversion
    • From: "Chris Chiasson" <chris@chiasson.name>

• References:
  • General--Mathematica and Subversion
    • From: daimos@sbox.TUGraz.at