At Sun, 18 Apr 2010 18:34:36 +0200, Jean-Yves Migeon
<jeanyves.migeon%free.fr@localhost> wrote:
Subject: Re: proplist
>
> Adding fuel to the discussion: I don't think that any particular format
> will fit everybody's taste.
Indeed, though taste (i.e. the "bikeshed" part of the argument) isn't
necessarily relevant.
To the extent that it is though, one could argue that in a unix/posix
environment the tradition would strongly favour line-oriented and
AWK-compatible table formats, perhaps extending to some C-like syntax
for more structured data.
> Concerning I-expressions, I find them
> particularly difficult to "follow" when you are submitting diffs.
I don't think I-expr formatted data is particularly more difficult
w.r.t. its use in contexts where diff's of the data might be exchanged
than S-expr, JSON, etc.
Yes, I-expr data will be more difficult to patch correctly whenever the
patch is applied to a different revision of the original than the diff
was generated from, but that problem can apply to any format if the
syntax-related formatting of the data isn't very carefully standardised.
This issue with diff/patch only matters of course if the format is used
to store data (rather than just for interchange) and where looking at
changes to this stored data makes sense and perhaps especially where
such changes to the stored data are expected to be shared in patch form.
> For example, diff -u on python scripts are sometimes a royal PITA
> (spaces vs tabs, different convention for indentation, the like),
Optimal change detection and always works better if the difference
algorithm takes into account the format and syntax of the input, _and_
if the syntax-related formatting of the data is very carefully
standardised. Neither of those conditions holds for Python scripts. ;-)
Diff and patch are designed to work best with unix style text and data,
i.e. line oriented records.
--
Greg A. Woods
Planix, Inc.
<woods%planix.com@localhost> +1 416 218 0099 http://www.planix.com/
Attachment:
pgpxy2qBsTgq_.pgp
Description: PGP signature