[Info-vax] FreeAXP/Avanti 2.2.0.435 Released

[David Froble]

Stephen Hoffman wrote:
> On 2012-10-17 23:16:03 +0000, David Froble said:
> 
>> JF Mezei wrote:
>>> On 12-10-17 17:10, VAXman- @SendSpamHere.ORG wrote:
>>>
>>>> I've ported 100s of VAX applications, utilities, drivers, etc. to 
>>>> Alpha.
>>>> I've never found anything difficult or impossible, and I never found 
>>>> a single bit of VAX software that was not portable.
>>>
>>> Try porting VAX applications that made use of Message Router callable
>>> interface.  Since the later was not ported to Alpha, it becomes prety
>>> hard to port your apps to Alpha.
> 
> Not the first prerequisite that's gone missing.  Undoubtedly not the 
> last, either.
> 
>> What is so hard about writing your own messaging software.
> 
> Well, beyond designing and writing and maintaining it, getting past my 
> own dislike of writing stuff that somebody else has probably already 
> written?

If it's already written, AND AVAILABLE, then it should be a simple 
matter to just "get it".

> It might pay to look at ZeroMQ, or another package, for instance.
> 
>> Might be more usable to your particular software than some generic 
>> package.
> 
> For me, it's more a case of what's central to the value-add I'm aiming 
> for, and what's not.  If there is some reason {performance, features, 
> whatever} to create a custom implementation, have at.  But for other 
> cases, rolling your own implementation ties up your own staff writing 
> and supporting something that doesn't move your primary {product, 
> service, whatever} forward.  That's bad.
> 
>> Can you give me examples of procedures than can be done on a VAX that 
>> cannot be done on an Alpha, or IA-64?
> 
> Accessing Q-bus and Unibus widgets would be somewhat gnarly.

Somewhat gnarly on my MicroVAX 3100 and VAXstation 4000 systems too ..

> You'd have to roll your own connect-to-interrupt support.
> 
> Code that's dependent on precise traps or that makes heavy use of 
> exceptions can get slow.
> 
> There are other snags, and there are other products that never made it 
> over.

I should have qualified my question a bit more ....

>> There is a reason some of us are entitled to the title "Software 
>> Engineer".
> 
> Pesky software engineers; get off my lawn!
> 
> But more seriously, there's a lot of writing of code yet to do, but 
> software integration is increasingly important part of the job.
> 
> And yes, sometimes you integrate on something that ends up getting 
> retired, or not ported, or not maintained.  Message Router, CDSA, etc.  
> Or dependent on something that's a moving target, such as OpenSSL.  
> Welcome to the other part of the job.
> 

Yeah, been there, done that, and still doing that ..

But, isn't that why we (used to) get paid the big bucks ??

:-)