To: John Tserkezis

 BL> Brenton has a smart trick. He uses an "information" screen to
 BL> distract you.

 JT> Sorta like '95 puts up that boring screen to hide the
 JT> config.sys & autoexec.bat?

  Tricks of the trade... but you have to choose your screen well.
Borland's Delphi uses a picture of a droopy-faced Athena. It annoyed
me so much I edited it out, along with the registration!

 BL> Buy another computer and network them.

 JT> I have five pooties on my desk at the momemt. I was thinking of
 JT> using the 386 to process the mail while the mailer handles the
 JT> file/mail transfers only. I couldn't be bothered really, the
 JT> load on the mailer at the moment is rather low, so I can't
 JT> warrant an extra machine to take that (little) load. 

  The cost is so slight that you might as well. A 386/40 is basically
worthless, but it's a fast machine! This was what I meant by the
"mature" phase of computer development. The step up to 386 left two
steps back (to 86 XT) with large performance drop, but the step to
Pentium 200 leaves the 386/40 as a very useful machine, still, and the
distance between a this-year's computer and one three-years old will
only shrink. You might as well have one a specific job like
communications.

 JT> I don't understand. I know what MS is trying to do with NT,
 JT> which is have a common OS where you can run the same software
 JT> on different hardware. That was the whole point wasn't it- make
 JT> software that is completely independant of hardware?

  What M$ is trying to do, is make lots of money. Hardware has grown
together, naturally. What we have is a few large software companies,
a few large CPU and chipset companies, and hundreds of small
assemblers. It *will* standardise.

  Software development (programming) is rushing ahead with standard
objects. Programmers are dealing in larger units of code. It seems to
me that hardware has to do something similar. Instead of processing at
the bit-level or 64-bit words, they have to make a large step up and
process kilo-bit objects (or something).


 BL> I hate to say it, knowing it'll set you off, but you have to
 BL> get comfortable with pointers to write in Windows, whether
 BL> Pascal *or* C, and once you do that, it doesn't matter which
 BL> language you use (except Pascal is easier to read and C writes
 BL> really horny shorthand). 

 JT> I can use pointers without stuffing around with windows. I've
 JT> only really played a bit with pointers so far. I was using them
 JT> to point to dynamically allocated memory after the main program
 JT> loads. It makes things a little more difficult for me, but
 JT> makes the software more flexible. 

  This is true... but I find now that I would *rather* use pointers,
in DOS Pascal too! In Delphi and Windows, you basically *have* to use
pointers and objects. You can *force* it to work the other way, which
means it's a good way to learn, but it's messy and it keeeps pushing
you towards using objects, and objects are *all* pointers. Pascal
itself is not like that. In fact, using objects in Pascal is not
*really* natural, but when you go to Delphin and Windows it is... and
you end up gaining the object-pointer mindset almost painlessly. It's
actually fun. You can do so much on the screen for so little code that
sheer laziness seduces you to objects. The idea of it grows on you.

  And from there, it is not a large step to the more awkward C++.  

 BL> I've written a million words on three fingers, I can do 40 wpm,
 BL> but I take your point.

 JT> I used to type about 30wpm with three fingers. Now I do about
 JT> 60, and could go faster if my hand-eye co-ordination was
 JT> better. I can't even wank proper. There is no way I would have
 JT> been able to do 60wpm with three fingers, not even if I used my
 JT> dick as the third finger.

  (chuckle) Yair. At 40wpm and two fingers it gets pretty desperate.

 JT> Dunno, I started touch typing after I got my hands on a pirate
 JT> copy of a touch-typing tutor program. :-) Maybe if I payed for
 JT> it, I might be able to type faster.

  ROFL!! It'd work for me! If I paid money, I'd have to get my money's
worth.

 JT> Regardless, you don't *have* to know pointers, but it certainly
 JT> helps, you can get a whole lot more done...

  The main trick for me, was to realise that you always have to have
*real* memory somewhere, for the pointer to point at. You can read a
file and load it into a buffer, and then just run pointers up and down
the buffer... but if you want to use the pointer somewhere else, then
you have to COPY the actual data to an actual memory somewhere else,
too, or the pointer ends up pointing at nothing much and runs off the
ned of the real memory. Once I realised I ALWAYS had to have real
memory, somewhere, pointers were fairly easy.

 BL> I meant that there will be a basic shift in the hardware
 BL> itself... perhaps a computer that handles standard objects or
 BL> tokens; a step *UP* in complexity that makes it externally
 BL> simpler.

 JT> You mean put part of the OS in the bios? Nope. That makes the
 JT> machine more expensive, with all that extra rom..

  ROM is very cheap, and RAM is cheap too, now. We use a standard
keyboard font, so why not take that much further and access objects
the software would use directly, both to paint the screen and run
mini-programs independently?

 JT> That and, the OS can overwrite some of the bios functions
 JT> anyway. As a matter of fact, that's what they do now anyway.
 JT> Full bios compatibility may have been of upmost importance a
 JT> while back when your "PC" may not have been fully
"compatible",
 JT> but nowadays this is not the case.

  Of course. You would still be able to overwrite these BIOS objects,
but you woudl also have the opportunity of using them lightning-fast
as complete, standard objects. It would need more than one CPU.

 BL> The same applies in commercial operations too. It would be very
 BL> rare if an office *needed* more than a 486/100 and Win31, or
 BL> more than 16-bit programs, butthat's not the way it works.

 JT> This depends. I don't see anyone upgrading unless they are
 JT> keeping up with the joneses, or the users complain about it
 JT> being too slow.

  (grin) The staff always complain it's too slow.

Regards,
Bob


 
___ Blue Wave/QWK v2.12
@EOT:

---