Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!usc!hacgate!ashtate!dbase!dukel
From: dukel@dbase.UUCP (Duke Luper)
Newsgroups: comp.databases
Subject: Re: dBASE IV - is this for real?
Message-ID: <136@dbase.UUCP>
Date: 5 Jul 89 00:12:41 GMT
References: <9383@xanth.cs.odu.edu> <135@dbase.UUCP> <802@forty2.UUCP>
Organization: Ashton Tate Devlopment Center Glendale, Calif.
Lines: 90

In article <802@forty2.UUCP>, eichi@forty2.UUCP (Stefan Eichenberger) writes:
> In article <135@dbase.UUCP> awd@dbase.UUCP (Alastair Dallas) writes:
> 
> >Undocumented trick of the month: invoke dBASE IV with a /t on the command
> >line to avoid the opening graphic (but not the legal message).
> 
> Ashton Tate seems to be busy documenting not yet implemented stuff rather
> than what people need to at least beeing able to start there program befor
> the coffee break. (I have been told by a AT representative at a local
> exhibition, that they only mention EMS support in the manuals (and adds?)
> but it's NOT YET implemented!!!!!!!!!(/&()&&/)%&/$%&]$&$$%&$!!!!!!!!)~

Nothing pisses me off more than someone who can't spell or write English and
can't follow a thought to its completion.  The above sentence about EMS
support has nothing to do with Alastair Dallas' comments.  So, why include
them (Alastair's comments) in this mail about dBASE?  Rag on dBASE all you
want.  I believe in free speech and have my own personal feelings about the
merits of dBASE IV (and Clipper, etc...).  By the way,  the phrase "pisses
me off" is a colloquialism and should not be construed or interpreted to be
"correct English".  :-)

> >Run some tests--try a DO WHILE loop, for instance.  Compiled object files
> >execute an order of magnitude faster than line-by-line.  Obviously, you're
> >still upset by the slowness of the opening graphic :-)
> 
> They still haven't learned, that a condition of a DO loop may contain
> a macro substitution. And the rule, whether it is allowed or not is so
> complicated, it fills about 15 lines in the manual. At least this bug,
> which persists ever since dBase II (not III) is documented.

What rule?  Macros?  He's talking about performance differences between 
dBASE III and dBASE IV.  Is this what you are addressing?  It doesn't seem
so.

> >If the object file is loaded into memory, it will be executed from memory.
> >If it is on disk, the time stamp of the object file is compared to the 
> >source program and it is automatically recompiled, if necessary.  If you
> >use the dBASE IV editor, the object file will be flushed, forcing a 
> >recompilation. You seem to be running your program, editing the source, 
> >rerunning it and expecting dBASE IV to know that you've changed something.
> >In the interest of speed, dBASE IV doesn't check the source file every time
> >it runs a program that is already in memory.  To force the recompile you
> >seem to want, issue a CLOSE PROCEDURE command before invoking your editor.
> 
> Ever heard of make? Or worked with Turbo Pascal Version 4, 5, 5.5? Turbo C
> Version 2? ...? Come on, AT, these programs can't be so expensive to buy and
> try with all the money you get from poor dBase IV buyers?

You seem to be comparing apples and oranges.  Turbo Pascal and Turbo C have a
combination of different executable files which each perform specific tasks. 
The powers that be at Ashton-Tate chose to implement the entirety of dBASE IV
into one executable file with many overlays.  This comes out to about 3.5 MB
worth of executables and overlays.  If Ashton-Tate had so chosen to break it
apart into different executable modules (like APGEN in one, SQL in another, 
etc...), then it would make it much easier to implement something similar to
a MAKE facility.  It makes it much more difficult to implement ANYTHING in
dBASE IV (including a MAKE utility) which runs in harmony with the rest of
the product.  Given that Ashton-Tate concentrated a lot of time on other
enhancements across the board, they apparently felt that a MAKE utility was
not as important as maybe SQL.  You can't please all the people all of the
time.  I am not defending their actions, but they have a lot of ears to
listen to in the user community and each person feels his/her desired addition
to the product is "the most important addition".  Remember, it's always easier
to get a few guys together and write an INCREDIBLE software package than it
is to hire a lot of people and try to coordinate a mammoth project.  dBASE IV
was certainly a mammoth project.  I do not disagree with you on the merits of
including a MAKE facility in dBASE IV.  It just comes down to what the
management at Ashton-Tate feels the user community wants the most.  PLEASE
don't start flooding comp.databases with what ASHTON-TATE should have placed
into dBASE IV.

SIDENOTE:  I am a consultant to Ashton-Tate working on the multi-user
features of dBASE.  I support the company and the product, even though there
are a lot of things internally I would like to see change for the better here.
My hope is that we can produce a MUCH cleaner and faster dBASE IV 1.1.  

DISCLAIMER:  Everything expressed in this mail item is solely my thoughts and
do not represent the thoughts or positions of Ashton-Tate or any of its
employees.

Duke Luper
Ashton-Tate
Torrance, CA

P.S.  I read a mail item earlier today from someone who said that sometimes
dBASE III would skip over an ELSE condition if the IF portion did not evaluate
to true.  If you read this mail item, could you include a listing of one of
your students programs?  I would like to look into this.  This is the first
occurrence I have heard of this happening.  Maybe it will help me flush out
an existing bug that we are yet unaware of?