Mr. Zack Urlocker of Borland wrote in Delphi Informant about the first
'marathon'. He said:

        "After we had finished Delphi 1, Gary Whizin, the Director of
Delphi Development, used to joke that it was like running a 10K race,
then discovering you had entered a marathon. It was certainly no sprint
to the finish, but rather, a lot of hard slogging. We set our goal and
the worked to achieve the milestones that would lead us to the goal.
We've found it to be a pretty good way to do things."
        "...we paced ourselves, slogged through the hard parts, and yes,
finished on time. I've shipped quite a few products in my career, but
Delphi 1 stands alone. It's the product I'm most proud of."

It was indeed an excellent race. With no hassle, with no potential
challenger. He went further stating how they had a head start with
the second race, to ship Delphi 2 within one year. What happened next
is now good history.

But there's one thing that caught my attention. It was the
editor's note, and I quote:

    "He's ", (referring to Zack), "currently working on the next Delphi
'marathon'"!              

What the heck is this next 'marathon'? Delphi 3? or Delphi 1.5?
Should they go on and maintain their lead? Should they go back on the
race track and wait for those slow challengers and fellow
Delphians? or both?

Whatever it is, I think the clamor now is for Borland to go back awhile
and have Delphi 1 restructured to have the full excellent features of
Delphi 2. Isn't it worthwhile for Delphi 1 to have visual form inheritance,
TQuery making use of dBase index files, data modules, TTable filter
property, enhanced DbGrid, Internet components, multiple inheritance,
built-in QuickReport, etc, etc..

Borland laid the bait, we got hooked. Now, it's time they should
uplift us more high by upgrading Delphi 1.

(***************
  Dante Galapon
****************)

I agree!  I think Delphi is fantastic, and I would love to move to
version 2 but the simple fact is that my customers arent ready yet so
I still need to develop with version 1 for the next year or two.  I
would sure appriciate a version 1.5 at this point.

But there are too many people, who must stay with D1 because
of Windows 3.11. So Borland should support also the more conservative
software users.

Ulrich Kamp

Ulrich Kamp
Haus der Geschichte
Adenauerallee 250
53113 Bonn
Tel. 0228 - 9165 - 130
FAX 0228 - 9165 - 302
mailto: k...@hdg.de

But there are too many people, who must stay with D1 because
of Windows 3.11. So Borland should support also the more conservative
software users.

Ulrich Kamp

Ulrich Kamp
Haus der Geschichte
Adenauerallee 250
53113 Bonn
Tel. 0228 - 9165 - 130
FAX 0228 - 9165 - 302
mailto: k...@hdg.de

And why put much effort into it? The underlying API sure as h--- ain't
changing! (as opposed to all the new APIs introduced on W95 & NT)

In fact, every time they expand the 16 bit product, they'll hit some ceiling.
D1 is already very resource hungry and cripples many developers who actually
run the development environment on the same platform they test on (Win 3.1x)!

Sure, they will probably release 1.5, all I'm saying is that the limits
imposed on Borland in the 16 bit world will not make it practical to
implement everything we'll see in the newer 2.0 (and soon 3.0) products.

--
=\                       homepage @ http://www.qsd.no/staff/rune
 *=- R.Moberg, CD-Player Pro info @ http://home.sn.no/home/mobergru/
=/               my own webserver @ http://pumba.qsd.no

Personally, I will eventually be interested in Latte.  I doubt that I
will ever be interested in a web page creator from Borland.

Bill Hunt

Rune Moberg <mober...@oslonett.no> wrote in article
<LGXVy0JfFewI08...@oslonett.no>...

How would D1 implement a D2 style 'String' type?  i.e. up to 4GB in size?!?

What *would* be nice is if D1 could have a String type which implemented
the
reference count idea, and perhaps had a maximum length of (say) 64K.

Similarly TThread could *not* be provided using system thread objects (they
simply don't exist in 16-bit Windows), but could they provide an internal
scheduler to simulate/emulate a threaded environment.  This may be too much
work given the future of a 16-bit Delphi (bound to be limited in the long
run), but
someone did observe that such a feature was implemented in previous Pascal
offerings from Borland, so the foundations for such a feature may already
be laid.
This would give Delphi a clear advantage over *any* 16-bit offering from
*any* other
tools vendor!

Just my 0.02.

Jolyon.