[lsb-discuss] Google Summer of Code

Wichmann, Mats D mats.d.wichmann at intel.com
Wed Mar 19 15:09:45 PDT 2008


Benjamin Kosnik wrote:
>>>> Would you mentor it, Mats? If not, anyone in our web
>>>> server/infrastructure team would do it?
>>> 
>>> no, I've got no clue about the software actually involved. I just
>>> know what I want.... 
>>> 
>> 
>> So perhaps put up the description of the problem and what you want to
>> have. If we attract students on this one, we can think later about
>> how we let a student implement this.
> 
> Yep. I see this has not been added to the wiki yet... can somebody do
> so? Or is it too late for SoC?
> 
>>> I guess this would actually be OpenJade... since this issue
>>> is with tools, not with the docbook definition itself.
>> 
>> Another question is: Is this really something new which requires to
>> be implemented or is this a bug in an existing feature. In the latter
>> case we cannot open a Google Summer of Code project on it. We must
>> report the bug to the appropriate free software project so that they
>> fix it.
> 
> Just who has the responsibility for fixing this is up for debate. Is
> it xsltproc? xsl stylesheets? docbook? openjade? xmlto? xmlroff, etc?
> Most likely, it's a combination of all of these.
> 
> This is precisely why an organization such as LF should take
> the initiative here, IMHO.
> 
> I strongly second this need. (libstdc++ docs were
> recently converted to docbook, where I ran into this. Similar issues
> exist for all FOSS projects using docbook, which include: Fedora,
> parts of kernel, GNOME, KDE, subversion, oprofile, valgrind, etc. etc
> etc.) I believe this is a significant issue across all platforms.
> 
> This is a real and significant issue in the free/open software
> community: on one hand, persistent naysayers point to the lack of
> docs, and on the other, developers who try to combat this then finding
> themselves working with inferior tools.
> 
> It would be great to see LF champion improvements on the print output
> side of the free documentation toolchain, even it this effort was
> separate from SoC.

I'm more than happy to describe the problem - and I don't think it's
too late to add something to the wiki.  As I noted, I'm unversed
in *all* the technologies (once we get past the docbook markup
itself) so I'm really hestiant to put myself down as a mentor.

I was hunting around the other day and saw this was on Fedora's
list of possible projects for the 2007 SoC, and somebody made
a proposal, but I guess from the lack of further hits that it
didn't go beyond that stage.  I actually kind of liked the idea
pitched there, since one of the issues has been that various
intermediate formats lose a lot of quality which contributes to
making the conversion to good pdf harder (for LSB, we've been
generating RTF and using Word+Acrobat on Windows to convert,
which has been good enough quality, even though I'm not happy
about the openness of the process, but now parts of the LSB
have gotten so big I can reliably crash Word on all the Win boxes
I have access to, so that's off the table as well!)

So how *do* we pursue this so it actually goes somewhere?

-- mats



More information about the lsb-discuss mailing list