[Date Prev][Date Next][Subject Prev][Subject Next][ Date Index][ Subject Index]

CHET G RE: XPL





XY-> Nope, II+. What was added? Nothing. (Does memory
 -> serve? Hmmmm. Okay, SU and GT, which are really just
 -> flavors of SV and PV, around the time of Xy3.10;
 -> and hmmmm extended permanent S/Gs 100-999 a little

What? Do you think XS insignificant?

I'm also partial to frame F. In terms of XPL, with the latter,
I wrote a program that could run any number of CIs on any number
of files. I still use that XPL.

I also wrote a series of self-teaching Xy XPLs on the basis of
frame F. Whatever its merit, it did make it possible.

 -> @RND, @REM -- whatever happened to them? And speed! (You
 -> think III+ was fast! I disagree with you anent IV, BTW; on
 -> strict benchmark numbers, yes, simple procedures are faster
 -> in III+; but the pithy terseness of the IV command set gives
 -> you so much more power, saves so much programming time, that
 -> going back to III+ would be a real deprivation -- IMO.)

One aspect of Xy, it never did follow a sense of normal
development, more of jumping here and there. About speed, in my
own timings, I found 3+ to be faster than 4 by a factor of 4. I
would say that it's primarily in terms of cursor moves within a
document that Xy has slowed down.

Moreover, I still haven't come across any functional task that
3+ couldn't handle. Of course, this could be ignorance, but
what's the alternative?

XPL in 4 is largely undocumented; the customization guide is not
particularly well written and most of the examples are lifted
from 3+. The two primary sources for learning XPL 4 are the DLGs
(XyDos is bad enough; XyWin is an adapted nightmare) and the XPLs
of you and Carl.

However, thanks to your count XPL, the pair of you present what
can be considered compiled XPLs rather than source code.
In any particular two XPLs the same material could have different
label names and different variable names. Thus you've done even a
better job than TTG of masking your work.

So XPL remains a true field of specialization; but unlike the
late 80s, it's hard to find the time to explore. In some ways
it's easier writing macros in WordPerfect because it has a
structured language.

 -> > A III+ solution, which I use since most of my Xy files are
 -> > ultimately for either Ventura or Quark, is to simply move
the
 -> > cursor CR while reading the cursor position before and after. If
 -> > a jump other than 1 or 2 occurs, CL and then linear right and
 -> > delete, since linear right doesn't "jump."

XY-> Sure, *if* you're eyeballing it and editing manually! (You
 -> can also use the JMP command, it goes right to the spot.)

Who's talking eyeballing? An XPL can do it much more efficiently;
and since the cursor moves are much faster in 3+ than 4, I
consider it time-efficient.

 -> These bugs are *major* hassles, and as often as not the
 -> workarounds open up new problems.

It's a hassle, sure, but anything unpredicted that behaves in a
predictable manner can be accounted for. Isn't that what
programming is about?

As far as theory goes, ever since the Renaissance there has been
the belief that things get better, which I believe is kind of
ridiculous; which you yourself have written about, i.e.,
psychiatrists (and the changeover of the DSM from personal
evaluation to shoot 'em with drugs). There's no particularly
reason why any new version is going to be better than the next,
since the theory is to gain new users as much as making the old
ones happier. You might as well blame market theory as anything
else.

But what killed personal computing was the network: After a brief
period of sunshine during which managers had lost control, once
more everything could not only be tracked but tracked as never
before.

The web pages are refreshing inasmuch as it's another burst of
creativity, but I wonder how long it'll last?

--Chet

---
 ? SLMR 2.1a ? chet.gottfried@xxxxxxxx  or  chet@xxxxxxxx