[lsb-discuss] ISV perspective in response to April 13 conf call

[Craig.Scott at csiro.au]

Considering the visibility aspect of the LSB, the LSB website is not really doing you guys any favours in terms of structure within the linuxfoundation site. On a whim, I pointed my browser at www.lsb.org to see if that would take me somewhere sane, but I got a fairly crude redirection message that took me to the LSB part of the linuxfoundation site. That's one set of points lost. When I got there, all the menu bar links, sidebar links and footer links were generic and pretty much not related to the LSB. There were a few links within the body of the page itself, but for something like the LSB which wants/needs visibility, this is starting the user from a less than ideal point of enquiry. There's a lot of screen real estate being used for things not related to the LSB.....

Regarding the "what are we missing" question, for us it is getting less about the technical impediments and more about the learning/resourcing. Writing code such that an app can be made LSB compliant is getting under control (for us it's now at the point where we rarely have issues on that front). Getting compliant RPM's built took a lot more work than it should have and to me, it still feels like the barrier to doing the things users expect (eg desktop integration, handling add/remove/startup of services, handling .deb-based systems) is too high for many people to bother. There are too many gotchas and quirky things that can go wrong. Some of these are just as hard on other platforms, but it is one area where we have lost much time in the past, often for what appears to be "nice to haves" rather than "die without them" features.

Having packages that are too old is another area that can catch us out, but we've essentially dealt with the majority of those by bundling our own more recent version of Qt with our packages. That has required working with Nokia to get patches in (work still ongoing). Most people won't go that far, we just decided to bite the bullet for ourselves and I had a personal interest. Most people will either work out how to not use/require the newer stuff in their own code, or more likely decide the LSB is too out of date to bother with.


On 20/04/2011, at 1:50 AM, Wichmann, Mats D wrote:

> 
> Craig (and John)... thanks for the notes.  
> 
> I thought they were very useful.  Both of the use cases
> are the sorts of things we /hope/ will work, but evidence
> seems to be there's way too little of that going on.
> Now if there were hundreds of such cases active, rather
> than... well, perhaps a handful...  the question
> continues to be, what are we missing?
> 

--
Dr Craig Scott
Computational Software Engineering Team Leader, CSIRO (CMIS)
Melbourne, Australia