Re: Future with Ada

["Ian S. Nelson" <nelsonis@earthlink.net>]

rob@z wrote:

> In article <9v57u1$mfb$1@nh.pace.co.uk>, "Marin says...
> 
>>Certainly, XML would be a good basis for reasons of portability. Not only
>>for GUI building but any sort of document formatting. So hypothetically, a
>>good start would be to build a word processor in Ada that stored its
>>documents in XML and, if possible, had its GUI built from XML.  
>>
> 
> you guys are going about it all the wrong way.
> 
> to make Ada more popular, you do not need to write yet another word 
> processor in it (star office is doing pretty well anyway).
> 
> Put the effort into reusable standard packages. there are so many
> areas where Ada has no standard packages to do things. Take Java 
> libraries as an example.


That's a good start.  It certainly hasn't hurt python or perl.  I think 
it's only part of it though.  I like ada and I see the benefits of it 
and it's easy to say what they are but with ada becoming part of the GNU 
compiler collection and possibly becoming more available to millions of 
hackers around the world it could be time to start putting money where 
the mouth is and produce a lot of free apps for Linux and BSD type 
systems.

For example, there are probably a dozen MTAs.  Security is a concern, 
reliability is a concern, performance is a concern and then the ability 
to shove the thing in to the various odd ball configurations people run 
for mail servers.  It's a hard problem and of those dozen MTAs, I don't 
see one written in ada.  DNS is the same way, there are probably 20 
different packages the serve up names.

I would think some kind of effort like the GNU project would be the best 
way.  Build a ton of tools and apps in ada and show that they are easier 
to maintain, more reliable (kind of hard since so many packages have 
been hammered on for decades) easier to extend and then ultimately 
better.  That's kind of what the AdaOS project is/was, except they were 
starting with the kernel and working out. It's a ton of work, it's not 
all terribly sexy work and some of it is difficult, I'm not sure how you 
build a better grep, but the end result should be well thought out tools 
and app that work, share lot's of common code and are easy to maintain 
and enhance.

Ian