home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Cream of the Crop 1
/
Cream of the Crop 1.iso
/
PROGRAM
/
TP020793.ARJ
/
02-07-93.TPC
Wrap
Text File
|
1993-02-07
|
40KB
|
1,131 lines
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-01-93 19:59:00
From Terry Hughes
To Michael Von_Kopp_Ostrows
Subject Async Pro
MV>Hi Terry,
MV>does Async Pro support compression routines (ZIP/LZH) too,
MV>or does it support only decompression of archieves.
Async Professional supports both compression and decompression of LHARC
1.13 files. It supports compression/decompression for ZIP 1.XX with the
exception of implode (it can decompress imploded files but it can't
implode them). We are currently working on adding imploding but I don't
have a release date yet.
MV>What's about "deflating" used in PKZIP 2.x ?
Nope, not yet. I think it's fair to say that our
compression/decompression tools will always lag the utility makers but a
good margin. We haven't even started thinking about deflating yet.
-Terry
___
X QMPro 1.0 41-2187 X TurboPower Software (voice 719-260-6641)
--- Maximus/2 2.01wb
* Origin: The Programmers Playhouse (1:128/60)
* Tossed by SFToss/286 v1.02a on 93/02/02 23:06:44
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-01-93 20:05:00
From Terry Hughes
To Louis Zarek
Subject TV browser for Filer
Louis,
Although the new TV browser is a bonus unit we won't be making it
available on our BBS right now. We only certify the new TV browser to
work with Filer 5.4 and beyond. And, anyone that has Filer 5.4 already
has the TV browser so there's no point in posting it on the BBS.
If you didn't get our latest mailer about new Filer features and how to
upgrade we probably don't have your correct address. You can call our
800 line for additional product or ordering information if you like
(800-333-4160).
-Terry
___
X QMPro 1.0 41-2187 X TurboPower Software (voice 719-260-6641)
--- Maximus/2 2.01wb
* Origin: The Programmers Playhouse (1:128/60)
* Tossed by SFToss/286 v1.02a on 93/02/02 23:06:44
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-01-93 08:00:00
From Dj Murdoch
To Mark Gryn
Subject PDN files (was: sorting a record)
-=> Quoting Trevor Carlsen to William Lagamba <=-
TC> Get the file TCSEL*.* from a PDN node or BBS near you and study the
MG> Okay, I've heard enough, any who is reasonably close to the Windsor
MG> (Canada) area, and has the latest TCSelect available for FREQ
MG> at high speeds, please give me a filename and an address...
I posted this the other day...
From: Bill Cassidy 249/99
By the way, if anyone want's to pickup the PDNPASCL stuff, they're available
from 1:249/1 @ 16.8k HST/DS and 1:249/50 @ 16.8k ZyXEL. Might be worth mentioni
g in the PASCAL echo?
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/03 09:46:29
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-01-93 08:02:00
From Dj Murdoch
To TOM MORGAN
Subject Re: OOP
TM> Are there any good books or tutors for OOP?
I don't know. I'm mostly self-taught.
TM> Also, why do some people program almost exclusively in OOP and others
TM> don't use it at all?
It's different from standard structured programming. Some people like it,
some don't.
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/03 09:46:29
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-01-93 08:06:00
From Dj Murdoch
To Etienne Boeziek
Subject Re: Tape
EB> I am looking for an tape unit/procedure.
EB> It must be compatible with a Jumbo tape (QIC-40 and/or QIC-80)...
EB> Can someone post this procedure/unit if it exists? Please...
EB> Or is there someone who can help me to write such a
EB> unit/procedure in TP7.0?
EB> Any help is welcome....
Apparently it's not easy. There are at least 3 different interfaces to the
Colorado Jumbo tape (the floppy disk interface, the FC-10, and the compressing
board). Each needs different interface code. I'd also suspect that a tape
drive from any other manufacturer would have a different interface.
QIC-40 and QIC-80 specify what's on the tape, not how to put it there.
Your best bet is to contact your tape drive manufacturer, or to call the standar
software with lots of EXEC calls.
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/03 09:46:29
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-01-93 22:50:00
From Dj Murdoch
To Scott Samet
Subject Re: more than 64k data
SS> If you restrict the size of the array elements to a power
SS> of two so that they never span a 64k segment boundary,
SS> it's not much trouble to GlobalAlloc a large
SS> block and do pointer arithmetic to locate a particular element.
That's a good point. In general, pointer arithmetic is messy, but when you're
working with powers of 2, things come out a lot neater.
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/03 09:46:29
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-01-93 22:55:00
From Dj Murdoch
To Scott Samet
Subject Re: Real Number Arithmetic
> Do you know the details about which tests were used?
SS> I have a mental block about statistics. As soon as anyone
SS> mentions the subject my eyes glaze over and I go into a
SS> trance. He explained it to me, and it went in one ear and
SS> out the other. The only thing that lingered was the part
SS> about counting the lengths of runs (increasing/decreasing
SS> sequences). He said they were far shorter than they should be.
I know Knuth published some tests for random number generators. Sometime
I'll run TP's through the tests, and let everyone know how it does.
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/03 09:46:29
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-01-93 23:57:00
From Dj Murdoch
To All
Subject BP Bug List
I keep a bug list for Borland Pascal (available from PDN nodes as BP7BUGS1.ZIP).
As I'm acutely aware, it contains errors - things which I've identified as
bugs, which just aren't. It's also missing bugs, since naturally enough I
don't hear about all of them or have the time to verify all the reports I
come across.
I think it would be much better for BP users if Borland published their own
internal bug list, as they currently do for BC++. I've told Borland this
before, and there have been some sympathetic listeners there; but still, it's
their policy not to publish any list of known BP bugs.
I've just received a message from Zack Urlocker, who is a manager at Borland
and closely involved with BP. He has asked me to get people who think they
would benefit from Borland publishing their bug list to write to him. If
you think you'd be better off seeing their list rather than mine, please do
write to him and tell him why. He'll be able to use your letters to convince
others at Borland that this should happen.
I expect mail to the Borland corporate headquarters would get to him (the
address is on the back of the manuals), or email to 76217,1053 on Compuserve.
I don't know of a Fidonet-Compuserve gateway, but if you can write to Internet
addresses, write to 76217.1053@compuserve.com.
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/03 09:46:29
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-02-93 20:32:00
From Trevor Carlsen
To Ruurd Pels
Subject Destructive B-Space?
RP> Quick trick to remove the last character from a string:
RP> if S[0] > 0 then Dec(S[0]);
^you missed the #
TeeCee
--- TC-ED v2.01
* Origin: The Pilbara's Pascal Centre (+61 91 732930) (3:690/644)
* Tossed by SFToss/286 v1.02a on 93/02/03 09:46:29
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-02-93 19:34:00
From Dj Murdoch
To Percy Wong
Subject Re: CONVERTING OLD TPU'S
PW> There is, nor will there be, a program to do this. It's how Borland
PW> makes money - it causes everyone to upgrade becuase no one knows the
PW> format of the files.
The format of the .TPU has been documented by William Peavy since at least
version 5.5. Borland makes money by selling compilers, not by selling .TPU
files.
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/04 20:39:25
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-03-93 08:34:00
From Terry Hughes
To Robert Jambor
Subject TurboPower Software
RJ>Hello Terry Hughes!
RJ> TH> We do very frequent updates of our products -- particularly when
RJ> TH> they're still in the early phase of their life cycle (as APRO
RJ> TH> is). Our products average 4 updates per year for the first
RJ> TH> two or three years of their life. With schedules like that
RJ> TH> it's pretty easy to fall behind in versions. But don't feel
RJ> TH> that you *have* to update all the time. Some people choose
RJ> TH> our Fast Update Plan so they can immediately get the latest
RJ> TH> and greatest; others only upgrade once a year or so or when
RJ> TH> some new feature compels them to.
RJ>So it seems. :-) What is this fast update plan?
Under the Fast Update Plan we immediately send out each minor release
(X.01, X.02, and so on) to you for one product for one year. If there is
a major update during your subscription period (like from 1.11 to 2.00)
you still have to pay for that but your Fast Update Plan continues for
additional minor updates of the new version.
Each minor update includes all bug fixes applied since the last minor
release (and a list so that you know what we changed) and, typically,
one or two new moderate to major features.
We just had a price increase in most of the Fast Update Plans; I believe
most of them are $80/year now. You can call our sales/info line,
800-333-4160, if you'd like more information.
-Terry
___
X QMPro 1.0 41-2187 X TurboPower Software (voice 719-260-6641)
--- Maximus/2 2.01wb
* Origin: The Programmers Playhouse (1:128/60)
* Tossed by SFToss/286 v1.02a on 93/02/05 09:16:56
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-03-93 08:57:00
From Terry Hughes
To Rob Kittredge
Subject Btree-filer
RK>Terry,
RK>TH>B-Tree Filer 5.4 includes a brand new TV-based browser. I think the new
RK>TH>browser requires 5.4 so it's not available as a standalone bonus unit.
RK>TH>You can upgrade to 5.4 for $49.
RK> I'm trying to decide whether to upgrade to 5.4..What features/fixes
RK>does it incorporate??
The major new features are BP7 protected mode support and a totally
automatic method of assigning workstation numbers. Note that the
protected mode support applies not just to Filer but to the supporting
units as well (SHARE, NETBIOS, etc.). Filer 5.4 also includes the new TV
browser mentioned above.
RK> I also have a comment to make about B-Tree Filer that may be of
RK>interest to you and others that use it. About 6 months ago I left you a
RK>message regarding the best way to process a FidoNet nodelist, adding
RK>records to B-Tree Filer. Someone suggested sorting the keys before
RK>adding them to B-Tree...Just a couple of days ago I started sorting the
RK>keys (400 at a time using MSORT) and it did reduce the add time slightly.
RK> I was also using the OOFILER unit on the bonus disk to make the record
RK>additions easier. I got to thinking though, using the OOFILER unit, for
RK>every record added, BP would have to do a virtual table lookup to get to
RK>my "createkey()" function...Doing this for EVERY record I thought was
RK>probably fairly time consuming. Boy was I right! I took out OOFILER and
RK>started using the procedural functions like BTAddRec, etc and it reduced
RK>my programs runtime from about 4 minutes to 45 seconds! On other, slower
RK>systems, it went from 45 minutes to about 3!
RK> You might want to pass this word on to others trying to get the maximum
RK>speed out of B-Tree...The OOFILER unit sure makes things easy but boy does
RK>it make things SLOW too!
If I had realized you were using OOPFILER to add all of your records I
would recommended against it immediately -- but not for the reason you
suggest since I'm not convinced that the virtual table lookup could add
that much time to the process. (A quick trip through the profiler would
tell you, though.)
A bigger problem is that using OOFILER would mean you were using logic
like this (some of it internal to OOFILER):
for I := 1 to MaxRecs do
for J := 1 to MaxKeyNumber do
BTAddKey(P, J, ...);
When adding batches of records/keys, though, it's much faster to do
this:
for I := 1 to MaxKeyNumber do
for J := 1 to MaxRecs do
BTAddKey(P, I, ...);
This results in fewer index page swaps during and, therefore, less time
to add all the keys. Perhaps you switched to logic like this when you
switched from OOFILER?
Also, while sorting the records ahead of time will decrease the time
required to add the records it may increase the overall time required to
add the records (sort time plus add time). Whether or not this is true
depends on your data and resources.
I thought I posted a similar message months ago when the topic first
came up but I guess you didn't get it.
-Terry
___
X QMPro 1.0 41-2187 X TurboPower Software (voice 719-260-6641)
--- Maximus/2 2.01wb
* Origin: The Programmers Playhouse (1:128/60)
* Tossed by SFToss/286 v1.02a on 93/02/05 09:16:56
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-03-93 09:02:00
From Terry Hughes
To Rob Kittredge
Subject Re: EXECWSWP question
RK>TH>Apparently, PK204C uses both DOS INT 21 and BIOS INT 10 to write to the
RK>TH>screen. Our EXECWIN and EXECWSWP units were only trapping INT 21.
RK>TH>New versions of both of these files should be on our BBS (719-260-9726)
RK>TH>by the time you get this message.
RK>Great! Thanks for letting me know!
RK>P.S. I just downloaded the new code and it still has problems. However,
RK>I've found that if I take out my DEVICE=ZANSI.SYS from the CONFIG.SYS
RK>file, the problem goes away. I could see the ZANSI causing problems,
RK>but if this were the case, the other program that uses PKZ 2.04 (SHEZ.EXE)
If you don't mind uploading a copy of ZANSI to our board (assuming
that's legal) I don't mind testing with it to see what the problem is.
RK>One more thing, I just got the update for PKZip 2.04e and in there
RK>there is an option to turn off "DOS fast character output" that PKZ
RK>appparently uses. I have no idea what interrupt the "DOS fast character
RK>output" function uses, but I'll be if you find it and trap for it in
RK>EXECWIN the problem will be solved. I'll look through some of my MSDOS
RK>interrupt listings and see if I can find it.
I'd say that "DOS fast character output" applies to some technique
they're using rather than a specific DOS function. That might refer to
calling int 10 directly rather than going through int 21.
-Terry
___
X QMPro 1.0 41-2187 X TurboPower Software (voice 719-260-6641)
--- Maximus/2 2.01wb
* Origin: The Programmers Playhouse (1:128/60)
* Tossed by SFToss/286 v1.02a on 93/02/05 09:16:56
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-03-93 09:17:00
From Terry Hughes
To Frank Derks
Subject FreePtr and FreeList
FD>The variable 'MaxAvail' holds the size of the largest contiguous free block
FD>_in_ the heap. The variable 'MemAvail' holds the sum of all free blocks is
FD>the heap.
One minor correction -- MaxAvail and MemAvail are functions, not
variables. That can be an important distinction when referencing those
functions in a time-critical section of code.
-Terry
___
X QMPro 1.0 41-2187 X TurboPower Software (voice 719-260-6641)
--- Maximus/2 2.01wb
* Origin: The Programmers Playhouse (1:128/60)
* Tossed by SFToss/286 v1.02a on 93/02/05 09:16:56
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-03-93 07:57:00
From Dj Murdoch
To Mike Copeland
Subject Re: Flush (was: Break & continue)
MC> I know this, and Dj used to moan about it whenever
MC> "standard" Pascal was referenced - the fact that TP
MC> doesn't conform to the Wirth standard.
No, I think that was Charles Falconer moaning. I think standard Pascal files
are bizarre; TP's file model makes a lot more sense to me.
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/05 09:16:58
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-03-93 08:11:00
From Dj Murdoch
To Kevin Higgins
Subject Re: Random Number Generators
KH> [[The following is yet another example of Greg Pierson's lunacy]]
GP> I'm looking for either a formula or the pascal source for a random
GP> number generator (evenly distributed) for a die rolling program. I
GP> need several different types of dice (6 sided, 8 sided, 4 sided, 10
GP> sided, 12 sided, and 20 sided). Any other information on the actual
GP> randomness of computer generated random numbers would
KH> also be of help.
KH> Here's some code for you:
KH> Function RndNorm(Mean, StanDev: Extended):Extended;
KH> Procedure RndNormalGenerate(var RndNormal1, RndNormal2 : Extended);
KH> Function RndNormal : Extended;
None of the ones you posted give even distributions. I'm not sure what's
wrong with the built-in Random(N) function; perhaps GP could tell us.
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/05 09:16:58
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-03-93 20:37:00
From Dj Murdoch
To Jud Mccranie
Subject Re: bp Bugs
DM> Yes, that's irritating, but I think it's by design. The calling
DM> sequences are marked in the help file in such a way as not to be
DM> reformatted.
JM> I think it is probably a bug. The calling seq. is probably the most
JM> important thing in the help file, yet they are chopped off unless you
JM> zoom the window. It simply doesn't make a bit of sense for it to not
JM> be wrapped in the window. The quality of the design at Borland has
JM> obviously droped tremendously in the last few years, but I don't
JM> think they would hire someone that bad at design.
It's not too hard for you to change this one - just expensive. Buy the Open
Architecture handbook, decompile the help files, and recompile them with the
syntax examples marked for reformatting.
JM> Do you send your bug list to Borland? A friend got a copy of BP 7 from
JM> Borland 2 weeks ago, and they haven't fixed any of the bugs yet.
It's a very big deal to release a new version. I wouldn't expect to see more
than a couple a year, so if they're going to come out with a bug fix, it probabl
won't be for a few more months.
JM> Another thing - in the IDE menu, for items that can't be selected, what
JM> is the point of even allowing the cursor to move to these items?
JM> Obviously the cursor should jump right over non-selectable items (a la
JM> Turbo Power's pick).
I think the point is consistency. To go down 5 lines, you hold the down arrow
down for enough time for 5 repeats. (Or hit it 5 times.) You don't need
to count how many items are grayed and how many aren't.
JM> Another one - if you don't have any windows open, you have to hit F10
JM> to get the cursr up to the menu line! This is a completely wasted and
JM> unnecessary keystroke. Futhermore, if you have another keystroke
JM> assigned to that operation (I have ESC for that) it only works if you
JM> are in a window.
I've never used F10. I use Alt-F, etc. to open the menu I want directly.
JM> Borland says that they're writing a major app in BP. I hope they're
JM> using the IDE and I hope the person has used 5.5 or earlier so they can
JM> actually see how frustrating and counter-productive these severe
JM> "design flaws" are.
Nobody at Borland wrote a major app in 5.5, or if they did, they kept it to
themselves. TP 6 was the first version which was actually used for one of
their commercial programs, as far as I know. I expect that's why there are
the minor tune-ups in the BP7 IDE - they noticed things were wrong, and fixed
them. However, the fact that the BP7 IDE is so much like the TP6 IDE means
that they like it. I wouldn't expect any movement back to a non-windowing
IDE like 5.5 in upcoming versions.
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/05 09:16:58
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-03-93 20:47:00
From Dj Murdoch
To Herb Brown
Subject Re: ISR's and objects
HB> Can an ISR be written into an object? For some strange reason
HB> me thinks not..
RP> Pray tell me why?
HB> An object gets confused if more than one instance is
HB> invoked. Crash and burn..
Hmm? I routinely have multiple TV windows open at the same time. Each is
an instance of a TWindow, but I don't get crashes because of that...
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/05 09:16:58
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-03-93 20:49:00
From Dj Murdoch
To Jud Mccranie
Subject Re: for Loop Control Vars
DM> It's illegal in TP, but the compiler doesn't check for it. The
DM> TP 6 manual used to make this clear: "It's an error if the
DM> contained statement alters the value of the control variable."
DM> The BP 7 manual is a lot more wimpy, just says your programs
JM> It has been that way since 2.0 (at least). If it is illegal, then
JM> why not make it a compiler error?
Because it's rather difficult to check. For example, is the following legal
or illegal?
procedure myproc(var i:integer);
begin
writeln('i=',i);
end;
var
i : integer;
begin
for i:=1 to 10 do
myproc(i);
end;
Here the variable i is unchanged, so it looks legal. But it's threatened
by that myproc procedure, which could just as easily have "i := 1" in it.
In order to decide whether this is okay or not the compiler would have to
know what myproc does, and TP isn't that smart. Alternatively, it could forbid
you from passing loop control variables as Var parameters, but that would
drive people nuts.
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/05 09:16:58
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-03-93 20:56:00
From Dj Murdoch
To David Mackintosh
Subject Re: Freq-ing
DM> I am a recent user preparing to make the big jump to "sysop" of
DM> my own board. In this area, I keep seeing "Freq this". I know
DM> that it means File Request, but how do I do it? The board
DM> I am using now has FIDO, GLOBE, and a couple others.
This is really the sort of question you should ask another sysop locally.
Once you get your software set up, it'll probably be easy to answer, but
until you do, it depends on all sorts of things.
None which has anything to do with Pascal, so you really shouldn't ask about
it here. Try to stay on topic.
The Moderator
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/05 09:16:58
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-03-93 20:59:00
From Dj Murdoch
To Jud Mccranie
Subject Re: real Number Arithmeti
JM> I think the sum of 2 uniform random variables is triangular. Summing
JM> more and more of them makes the distrubution closer and closer to
JM> a normal (bell-shaped) distribution. Summing 6 or more is very
JM> close to normal.
JM> Normal := random + random + random + random + random + random;
That's not a very good approximation, nor is it necessary. If you really
want a normal distribution, look up a good generator in any of lots of books.
A couple were posted here just yesterday; an equivalent one is below.
The problems with the sum of uniforms are:
- the range is wrong. Your sum has a range of 0 to 6; a true normal should
be able to take on any real number.
- the mean is 3 and the variance is 1/2. Since most people specify a standard
normal with mean 0 and variance 1, you need to do corrections to get a standard one.
- it's got a pretty chunky distribution (a polygon), while the normal should
be smooth.
Here's the one I use:
function rnorm(mean,variance:float):float;
var
t1,t2,r:float;
begin
repeat
t1 := 1 - 2*runif;
t2 := 1 - 2*runif;
r := sqr(t1)+sqr(t2);
until (r < 1) and (r > 0);
rnorm := mean + t1*sqrt(-2*variance*ln(r)/r);
end;
I'm using my own "runif" here rather than the TP one, but you could substitute
"random" without doing much harm. If you want to speed this up a bit, you
can actually get two independent random numbers at once:
rnorm2 := mean + t2*sqrt(-2*variance*ln(r)/r);
is independent of the rnorm value.
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/05 09:16:58
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-04-93 06:16:00
From Trevor Carlsen
To Percy Wong
Subject CONVERTING OLD TPU'S
-> Hmm..has anyone seen a program that can convert an old TPU into one
-> of the newer version of Pascal's TPU..Hmm..If so..can U please tell
-> me..Thanx...
PW> There is, nor will there be, a program to do this. It's how Borland
PW> makes money - it causes everyone to upgrade becuase no one knows the
PW> format of the files.
It is comments like the above that indicate a total lack of understanding
of just what it is that TPUs do and how they are used by the compiler.
The incompatability across versions has nothing to do with that and everything
to do with linking, entry points of routines etc etc. TPUs, and the specialised
information they contain, are the reason for TP code being so compact.
TeeCee
--- TC-ED v2.01
* Origin: The Pilbara's Pascal Centre (+61 91 732930) (3:690/644)
* Tossed by SFToss/286 v1.02a on 93/02/05 09:16:59
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-04-93 18:10:00
From Terry Hughes
To Dennis Alexander
Subject Async Pro Question..
DA> I have a question about AsyncPro... is it door routines or is it jus
DA>the low-level stuff? Do you have door routines? I'm having a hard time
DA>finding routines that fit my needs 100%... BOI is about perfect for my needs
DA>but the windowing isn't working for me. Do you include source code?
I wouldn't call it low-level since it provides routines from simple
serial I/O all the way up to easy-to-use protocol, modem and terminal
window routines. But, no, it does not have any door-specific routines.
Yes, it includes complete source code.
-Terry
___
X QMPro 1.0 41-2187 X TurboPower Software (voice 719-260-6641)
--- Maximus/2 2.01wb
* Origin: The Programmers Playhouse (1:128/60)
* Tossed by SFToss/286 v1.02a on 93/02/05 09:17:05
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-04-93 07:37:00
From Dj Murdoch
To Steve Connet
Subject Re: pointers
SC> TYPE MyString = String;
SC> MyArray = Array [1..10] of ^MyString;
SC> My2Array = Array [1..10] of MyString;
SC> VAR MyVar : MyArray;
SC> My2Var : ^My2Array;
SC> Now MyVar is an array of pointers, and My2Var is merely pointing to an
SC> array. But which is more "appropriate" to use? And under what
SC> circumstances would you use one over the other?
MyArray is usually the one you'd want. If you use NewStr and DisposeStr to
allocate/deallocate the elements, then it can be much more efficient - holding
10 strings of 23 chars each will take just 10*(4 bytes per pointer + 24 bytes
per string) = 280 bytes. My2Array will take 4 + 10*256 = 2564 bytes to hold
the same information.
One place where you might want a pointer is if you want this to be a function
result. Functions can't return arrays, but they can allocate arrays, and
return pointers to them. You might want a pointer to an array of pointers
to get the memory savings, though.
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/06 09:43:27
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-04-93 07:52:00
From Dj Murdoch
To Herb Brown
Subject Re: VMT
HB> Totally unrelated to the above topic, but, have you ever
HB> written a VMT simulation? I found an excellent example in
HB> a book by Namir Shamas. He uses an example to search a
HB> virtual memory array. I liken it more to a memory cache
HB> the way it's used, though..
I'm not sure what you mean by a VMT simulation, but it sounds interesting.
Could you explain a bit more?
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/06 09:43:27
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-04-93 07:57:00
From Dj Murdoch
To Arthur van der Peijl
Subject Re: Compression, Streams & at Random
AV> With PHuffmanFilter(Source)^ do
AV> Begin
AV> Writeln(GetPos);
AV> Seek(nr*73);
AV> Read(test,73);
AV> End;
AV> This program should read the recordnumbers of the
AV> compressed files BOSAL.LZH This file contains record with
AV> each 73 character, it is compressed via the huffman-program of Stream13.*
AV> My Program: I can't read the records! Each time i give
AV> Seek an order, he move the pointer 73 positions. It
AV> doesn't start at the zero position to seek (which it
AV> should do according to the docs -> because of compression,
AV> it must seek from position 0 on).
Looks like an error in the documentation. I never implemented the Seek for
this filter. You could add it in yourself, but it would be pretty slow.
It should be something like this:
procedure THuffmanFilter.Seek(pos:longint);
var
buffer : array[1..1024] of byte;
begin
inherited Seek(0); { or TBitFilter.Seek(0) in TP 6 }
while pos > 1024 do
begin
Read(buffer,1024);
dec(pos,1024);
end;
if pos > 0 then
Read(buffer,pos);
end;
This doesn't have error checking, and I haven't tested it, but it looks about
right.
A faster way to do this would be to force byte alignment after each record,
and keep a record of the starting points in uncompressed bytes. Then you
wouldn't need to seek back to the beginning each time.
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/06 09:43:27
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-04-93 08:06:00
From Dj Murdoch
To Diederik Beltman
Subject Re: geheugenprobleem Borland Pascal 7.0
DB> Onlangs heb ik BORLAND PASCAL WITH OBJECTS 7.0 gekocht en ik
DB> krijg het niet voor elkaar om BP vanuit DOS te starten. Als ik
DB> HIMEM.SYS uit m'n CONFIG.SYS haal lukt het wel, maar dan kan
DB> ik Windows 3.1 niet meer starten. Kunt u mij helpen het
DB> systeem op een juiste wijze te configureren? Ik draai onder
DB> DOS 5.0. De CONFIG.SYS en AUTOEXEC.BAT druk ik hier onder af.
DB> Bij voorbaat dank.
DB> Met vriendelijke groeten, Diederik Beltman.
This is an English language echo. I can guess that you can't get BP 7 to
run with your CONFIG.SYS setup, but I can't follow well enough to see your
exact problem. Post again in English and I'll try to help. It might be better
to call Borland about this one, though.
Duncan Murdoch
Moderator
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/06 09:43:27
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-04-93 08:09:00
From Dj Murdoch
To Matt Kracht
Subject Re: big linked lists
MK> I want to make a file viewer that will view files as
MK> large as the nodelist. I can make smaller file viewers
MK> fine, but don't have much idea on where to start with
MK> making 'em any bigger.
I wouldn't use a linked list for that. You probably don't have enough memory
to keep the whole file in memory at once, so don't even try. Just keep it
as an open file, and seek back and forth to read new parts. It'll go a lot
faster with a big buffer; a TBufStream would be ideal.
--- Msg V3.2
* Origin: Murdoch's_Point - - (1:249/99.5)
* Tossed by SFToss/286 v1.02a on 93/02/06 09:43:27
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-05-93 04:12:00
From Trevor Carlsen
To Luke Kinberg
Subject Integer Variables
LK> Is there a way to type in a letter or another non numeric
LK> character in an Integer variable without the programming
LK> aborting.
var
i : integer;
begin
repeat
write('Enter any number: ');
{$I- <<<<<< This turns off I/O error checking at runtime. }
readln(i);
{$I+ <<<<<< and this turn it back on again }
until IOResult = 0;
end.
That will do what you want. However it may be better if you handled the input
yourself using readKey.
TeeCee
--- TC-ED v2.01
* Origin: The Pilbara's Pascal Centre (+61 91 732930) (3:690/644)
* Tossed by SFToss/286 v1.02a on 93/02/06 09:43:31
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-05-93 04:24:00
From Trevor Carlsen
To Jakob Paikin
Subject output
SB> How do I make it so that only the things I want redirected
SB> will be redirected, while everything else shows up ONLY on
SB> the screen (no graphics, just TEXT stuff).
JP> Just put the line:
JP> DirectVideo := False;
JP> before the lines doing things you want redirected, and:
JP> DirectVideo := True;
JP> before lines that should not be redirected.
Why don't you TRY it before giving the advice?
HINT: It will not work because the advice is wrong.
To get redirection when the CRT unit is in use requires that stdout be reassigne
. (See the first page in the chapter on the CRT unit in the manual.)
TeeCee
--- TC-ED v2.01
* Origin: The Pilbara's Pascal Centre (+61 91 732930) (3:690/644)
* Tossed by SFToss/286 v1.02a on 93/02/06 09:43:31
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-05-93 04:32:00
From Trevor Carlsen
To Nico Keirsmakers
Subject ISAM
NK> I'm trying to make an index-sequential file to speed things
NK> up when retrieving data from my files. So I thought of
NK> creating an index-file with the record key (e.g. Name) and
NK> the position of that record.
NK> Now when I need let's say information about 'Mr. JANSSENS' I
NK> scan my index-file for Name = 'Mr. JANSSENS' and when it's
NK> found I get the position of that record in the master file
NK> (=pos). Then I open my master file and do a seek(master,
NK> pos) followed by a read of the specified record.
NK> I though this should have speeded up things, but in fact
NK> there's only a minor increase of the speed...
NK> Has anyone ever tried doing something like this ?? Cause I
NK> need to access this file rather quickly, is there any way of
NK> doing this really fast ?? I used reads and no blockreads
NK> because my file is record organised.
NK> [ When making ISAM files with Cobol, access is almost
NK> instantly, I never waited more than 1 second for
NK> retrieving a record and I tested with 50.000 records. When I
NK> did this in Pascal I had to wait for almost 15 seconds... ]
50000 names in the file means that you cannot load the index file completely
into memory. So one way to do this is to create an index array to your index
file and rather than create another file for this array, store it at the beginni
g of the index file and load it into memory at the initialisation stage.
type
idxArray = array['A'..'Z'] of longint;
Each element can contain the file offset of where each letter of the alphabet
starts in the index array. Then you can load just that section of the index
file into memory (using BlockRead of course), search it and then proceed to
the appropriate record in the master file.
TeeCee
--- TC-ED v2.01
* Origin: The Pilbara's Pascal Centre (+61 91 732930) (3:690/644)
* Tossed by SFToss/286 v1.02a on 93/02/06 09:43:31
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
Conference 4
Date 02-04-93 22:05:00
From Norbert Igl
To Luke Kinberg
Subject Integer Variables
Hello Luke!
03 Februar 1993 12:37, Luke Kinberg wrote to All:
LK> Is there a way to type in a letter or another non numeric character in an
LK> Integer variable without the programming aborting. Code would be
LK> appreciated. Thanks Luke
Function ReadInteger( message: String; x,y : Byte ) : Integer;
var I, result : integer;
begin
gotoxy( x,y );
write( message,':' );
x := wherex -5 ;
repeat
i := 0;
gotoxy( x,y );
write('_____');
gotoxy( x,y );
{$I- }
readln( i );
{$I+}
Result := IoResult;
if Result <> 0
then write( #7 ); { error-beep }
until Result = 0;
ReadInteger := I
end;
Norbert
--- GoldEd 2.40p/FD2.02/FastEcho
* Origin: Fly like an Igl (2:2402/300.3)
* Tossed by SFToss/286 v1.02a on 93/02/06 09:43:31