Proposal to use meinproc for Kubuntu

[matthew.east.ubuntu at breathe.com]

Sean Wheller writes: 

> On Wednesday 23 November 2005 11:44, Bhuvaneswaran wrote:
>> I've prepared a following proposal to change our build system for Kubuntu:
>> https://wiki.ubuntu.com/DocteamKubuntu
> 
> This proposal makes absolute sense and was the intended furture direction. The 
> only reason docs adopted xsltproc was because it was part of GNOME and it was 
> easier to get people who were working on the GNOME Ubuntu, to use the GNOME 
> toolchain. 
> 
> When Kubuntu was added to the doc portfolio it was just easy or a reduced 
> learning curve. 
> 
> However, now that more people are working on Kubuntu I would suggest a move 
> toward mienproc since it would mean better integration with the build 
> environment for kde and kde-docs in general. While doing this I would also 
> recommend reverting translation tools, for Kubuntu, back to those supplied by 
> kdesdk

I still don't see _why_ integration with the kde-docs build environment is 
necessary. AFAICS the end result is identical. Problems I see with adopting 
meinproc for kubuntu and xsltproc for gnome (partly repetition of previous 
email): 

* doing some docs with xincludes, and others without will increase the entry 
barrier and confusion for contributors who wish to work on both ubuntu and 
kubuntu
* having different build systems will make for more difficult and greater 
maintenance
* generally, fragmenting the team effort without an identifiable reason 
(although I will be pleased if someone points this out) 

In fact, I think it is more important that integration between ubuntu and 
kubuntu documentation is achieved, rather than integration with upstream. We 
are limited to a certain extent by the readers, but I don't see why our 
online documentation should use different technology. 

As for translation, the idea of using two different translation build 
systems I really find just baffling, especially since I know full well that 
we are short staffed and will probably not have the time to grapple with 
separate systems (the past two releases have taught me that). Again, I see 
no good reason. 

Matt