Problems with X.org and incompatibilities with in-house software

[Richard Brown]

Daniel Stone wrote:
> Hi Richard,
>
> On Sun, Feb 28, 2010 at 05:48:50PM -0500, Richard Brown wrote:
>   
>> To our much dismay we have recently found after attempting to install  
>> new Linux boxes that these extensions no longer appear to be available.  
>> This has caused most of our internal applications to blow up and to be  
>> completely ruined and unusable in the process. Dozens of applications  
>> have now blown up and are not able to be used, involving millions of  
>> lines of code. Thousands of dollars already invested in upgrade to new  
>> Linux systems appears to be completely useless now, as none of our  
>> applications can be used on these new systems.
>>     
>
> We advertised the deprecations fairly widely, and it was done in a
> staggered manner.  No-one complained.
>
>   
>> It is well past time that your organisation make backwards compatibility  
>> with core X11 and all extensions to it a primary principle of your  
>> organisation. To many have invested too much money into developing  
>> software to utilise these extensions than to have them mindlessly  
>> removed and thus blowing up dozens of our internal applications.
>>
>> We have decided that we will probably move to an entirely Win32 platform  
>> instead of investing hundreds of thousands of dollars into an extensive  
>> rewriting of our existing X applications, as it seems like, from what we  
>> have already seen, it no longer seems as though we can count on this  
>> platform to provide the backwards compatibility we need. We have been  
>> talking to Microsoft extensively about this issue and they have indeed  
>> provided us with huge resources and have iron clad commitments to  
>> maintaining compatibility with their older interfaces, so we can rest  
>> assured that with them that code we write today will still work years  
>> and years from now.
>>     
>
> The key here is in what you've said -- you're paying Microsoft to make
> these things happen.  If you want to pay a competent UNIX vendor (Red
> Hat or Sun, basically), then I'm sure they could help you guys out.
>
> It's pretty hard to do anything when you have to guess as to what
> someone who you've never previously interacted with is going to say in
> several years' time (when did PEX and XIE disappear -- late 1990s? early
> 2000s?).  But I'm pleased to announce that old versions of X still work
> every bit as well now as they do previously, so if you don't want to
> upgrade then no-one's holding a gun to your head.
>
> Cheers,
> Daniel
>
> (PS: Ask Microsoft how you go running 16-bit DOS applications under
>      Windows 7.)
>   
Thank you for your help, on this matter. I do apologise for the tone of 
the previous letter. I had just gotten back from a bar and was rather 
drunk. Anyhoo, I am not of asking X.org to maintain these old extensions 
which have functionality available elsewhere or no longer provide 
functionality, and have broken or questionable code security wise.

We will be staying with X.org. I am very high up in the decision process 
at our corporation (though i do try to avoid making decisions while i am 
drunk), so that decision is quite final. Our applications are not too 
dependent on these extensions and we are planning to update to fix our 
apps to not use them and keep them running.

I apologise if the previous letter was ill conceived. Many of my 
statements were made in poor judgment and were ignorant. I will admit 
that. (one too many of the cold  ones in one evening can do that, i need 
to break my drinking habit and attend my AA sessions more frequently).

I do think that x.org is a wonderful product and i do appreciate it. I 
do appreciate backwards compatibility in the core protocol and the 
network transparency is very useful to us.

I am interesting in helping with X development, but i do not know the 
sligthest thing about the X server, any pointers to explanation of X 
internals?

Thank you for the great window system and keep up the good work.

Sincerely,
Richard Brown