Hi, Paul.

PE> PE> Where can I FREQ it from?  If it's not FREQable, it's vaporware,
PE> PE> just like OS/2 4.0.

PE> FM> Interesting definition of "vapourware", on that
basis Corel Draw is
PE> FM> vapourware - you can't freq that.

PE> Or buy it in a shop.

What about the latest program you or your colleagues wrote at work? It's
not freqable, you can't buy it in a shop - vapourware? Interesting
definition.

PE> FM> What do you want me to do? Upload it
PE> FM> to TML? Why? You wouldn't use it, you don't even like it. And it's in

PE> I would use it once actually.

Oh, what for? And does that mean you want me to upload it?

PE> But it's not for me, it's for the
PE> brain-dead the need to combine packets.

I know what it's *for*, I wrote it!

PE> FM> Pascal so you couldn't recompile it if you didn't like the way it

PE> I have a 100% ISO conforming Pascal compiler here.

Probably won't help you. You usually write in ISO C, others write in K&R
C or whatever variations their compiler supports which is useful.
They're still writing in C.

PE> FM> doesn't do some things (there are at at least 2 documented, possibly
PE> FM> others not, design decisions which could well be not how someone else
PE> FM> would want it to act). Oh, and it doesn't have an "explicit public
PE> FM> domain message".

PE> It needs to have something if it is to reside on my BBS.

PE> PE> You're probably running an old version of PKTJOIN, it now has

PE> FM> I'm *running* my version of PktJoin. The PQWK260 archive from which I
PE> FM> have been examining the source to yours is dated 22/5/95.

PE> PKTJOIN is in PQWK260?  I don't think so.

Quite right, my mistake. The PKTJOIN archive from which I have been
examining the source to yours is dated 22/3/95.

PE> PE> error checking on the writing, and there have been NO reported
PE> PE> bugs.  And don't bother trying to say that misuse of the program
PE> PE> counts as a bug.

PE> FM> Robust software responds gracefully to misuse.

PE> You can say it is not robust, you cannot say it has a bug in it.

Perhaps you can say that, I'd call it a bug.

PE> PE> Some people want to run it on OS/2 32-bit.  Well, if they were
PE> PE> going to run it at all.  Linux is another requirement for any
PE> PE> software I use in the long term, as that's where I want to end
PE> PE> up going.

PE> FM> Then they/you can run yours, and take the risk.

PE> Take the only choice, actually.

Possibly, in that case. I have different choices.

PE> FM> Consider it non-existent then, if you want to propose that as a
PE> FM> criterion for existence.

PE> I do.

It's farcical, as my comment at the top points out.

PE> PE> FM> And as well as having at least those 2 fewer bugs,

PE> PE> Likely a complete lie.

PE> FM> I think we've discussed one before, the other one (or two, maybe the
PE> FM> same bug) I identified while coding mine and checking to see how you'd
PE> FM> implemented some possible failure modes. You hadn't.

PE> That's robustness.  The original PKTJOIN didn't check a single
PE> error at all.

Glad to see you fixed some of the bugs.

PE> PE> FM> it's 30% faster than yours

PE> PE> It may well be 30% faster than the one you have, but there is no
PE> PE> way in the world that it would be faster than the one I could
PE> PE> write if I had wanted to.  I would do so, except I don't have your
PE> PE> vaporware one available to keep you honest.

I didn't write mine especially for speed, just as it came out
"naturally" (this is Pascal after all). But as it happens I can't
off-hand think of anything which would make it *significantly* faster,
ie worth the bother.

PE> FM> Well create some PKTs which you think are big enough to fairly test it
PE> FM> and run your best version on them. Let me freq them and I'll run mine.

PE> I want to see yours to keep you honest.

I am honest. I told you that.

PE> FM> I'll certainly believe your times, why would you want to lie in an
PE> FM> objective evaluation? Or do you *really* want me to upload it?

PE> Yep, if you want a real test, you have to do that.  First I will
PE> be checking to see that it was written in Pascal, not assembler.

Not a line of assembler in it. And it compiles with my Pascal compiler.

PE> PE> FM> and both the source and executable are smaller.

PE> PE> That I can't be sure of, only the speed.

PE> FM> Yes you can, I just told you.

PE> You can't be sure of what I can write if I actually have some
PE> performance/size objective I am trying to meet.

Sure, I wasn't talking about what you could write, just what you did.
You could probably write a bug-free one too.

PE> PE> Vaporware you write.  Yeah, between your PKTJOIN and Bob's
PE> PE> PKT2QWK, you've got the most impressive collection of vaporware
PE> PE> I've ever NOT seen yet.  BFN.  Paul.

PE> FM> It exists,

PE> It doesn't, it's vaporware.  Just like OS/2 4.0.  No point comparing
PE> OS/2 4.0 to Windows NT and saying OS/2 is the better OS.  It isn't.
PE> It doesn't exist.  BFN.  Paul.

I've never seen it either. But then I don't think I've seen any version
of OS/2, or WinNT. But I have seen my PKTJOIN, it exists.

So do you want it or not?

Regards, fIM.

 * * I have a firm grip on reality. Now I can strangle it.
@EOT:

---