[Info-vax] September 6, 2016 - new Roadmap and State of the Port updates now on VSI website

[David Froble]

Paul Sture wrote:
> On 2016-09-10, David Froble <davef at tsoft-inc.com> wrote:
>> johnwallace4 at yahoo.co.uk wrote:
> 
>>> I guess you weren't a fan of VMS POSIX then?
>> Nope!
>>
>>> If it had lasted, if the customers who said they wanted open standards
>>> had walked the walk as well as talked the talk, many of the "porting"
>>> and "tools" issues being discussed round here would have been non-issues.
>> If what one wants is *ix, then why not just use *ix?
> 
> Unfortunately there are lots of *nix goodies which are quite useful, and
> these tend to be well tested, with most corner or edge cases addressed.

Yes, and a FTL Starship would also be a nice thing.  Today, not in the future. 
But we can't have everything, at least without some form of "purchase" now, can we?

> Port or reinvent the wheel?  Your choice of course, and given your
> aversion to C, which I do understand, there can be a solid case for
> rewriting into another language for maintainability, but then you have
> got something you have to maintain yourself.

Unless it's compile, link, and run, any port is in some respects "reinvent the 
wheel".  Once you make changes, as the original changes with time, you must 
re-implement your changes, and sometimes they may not be compatible.

> And during your rewrite, did you manage to address existing security
> holes in the original product?  While a simple rewrite into another
> language should eliminate typical C gotchas, it won't address things
> like the version of SSL you are calling being retired due to security
> issues.
> 

Oh, yeah, punch my magic button!

:-)

I agree, but that's something you're going to have to address anyway.

If people want to port *ix software, fine, no problem.  But, when I get a bit 
testy is when they start to blame any issues on VMS, and not on the original 
product that would appear to have been written without any concern for portability.