[comp.sys.mac] Information, please

mikes@tekecs.TEK.COM (Michael Sellers) (01/14/87)

I am considering investing in a C-type language for my Mac (512K, 400K
disk drives, no HD, old ROMS -- is this the new "vanilla"/"skinny"/deprived
Mac? :-[ ).  What is the best on the market these days?  Is it still
Lightspeed-C?  Are there any (good) C++ or Objective C compilers out there?
What am I looking at in terms of price/performance?  Any help, pointers,
recommendations, or testamonials that I could get would be appreciated greatly.
-- 
			   Mike Sellers
     UUCP: {...your spinal column here...}!tektronix!tekecs!mikes

	 	       Otium cum Dignitatum

joel@gould9.UUCP (Joel West) (01/15/87)

Objective C is vaporware.
C++ ports have been done, probably none officially, and Apple will
be supporting something like C++ with MacApp in about 12 months.

MPW C is for large systems, reliable, bulletproof.
Lightspeed C is quick and easy to use.

Consulair is still around but I have no experience with it.

-- 
	Joel West			     MCI Mail: 282-8879
	Western Software Technology, POB 2733, Vista, CA  92083
	{cbosgd, ihnp4, pyramid, sdcsvax, ucla-cs} !gould9!joel
	joel%gould9.uucp@NOSC.ARPA

saf@clyde.UUCP (01/16/87)

I am confused by the differences between the Mac Development System (MDS) and
the Mac Programmers Workshop (MWP). Apparently, MDS has been around a while
and is mature (not that I have been able to find a copy except as a special
order).  MWP appears to be in Beta test and is available from APDA (and maybe
others).  

Which (if either) should I buy?  I am hesitant to shell out $100 for MWP only
to find that Beta=bugs galore.  Should I wait for it to become a released
product or will I be waiting for a year or more?  Does it look like MWP C
will become the standard and displace LightSpeed?  Finally, I wonder what the
released price will be as compared to the Beta price...

I would appreciate any info available on the quality/value of these products.

	Steve Falco {clyde, floyd, bonnie, moss}!saf

mce@tc.fluke.COM (Brian McElhinney) (01/21/87)

MPW was announced as being out of beta during the MacWorld expo, but
the new manuals had not yet been printed.  The only source is ADPA.
If you want a copy, you should get your order in to ADPA, as they will
likely be swamped (as usual).

MPW is an integrated development environment (with, unfortunately, a
new regular expression syntax).  The main advantage to MPW is that you
can use Object Pascal and MacApp.  Object Pascal is an extended Pascal
(blessed by Wirth) with messages, but no other object-oriented
extensions.  MacApp is the ultimate empty Macintosh application, and
is written in Object Pascal.  It has been estimated that, once
learned, Object Pascal and MacApp enable a finished application to be
developed in one-fourth the time of traditional procedural languages.

Object Pascal and MacApp are sold separately from MPW, so the total
cost is more like $300-400 (I didn't pay attention to the release
prices; well worth it in my estimation).  There is also a C compiler
and Object Assembler, and all three languages can be mixed (though C
has no way to receive messages; but you can call C from an Object
Pascal method).

I have some problems with the lack other object-oriented extensions to
Object Pascal.  The official line from Apple is that it confuses the
poor programmer.  I feel that learning to program based on data
objects that can respond to methods is what slows most programmers
down, and obviously Object Pascal still requires that.  Once this is
mastered, the more "advanced" concepts can be easily learned.  After
all, aren't programmers supposed to be expert users?

Eventually, you may be able to program in Object Pascal without MPW (I
understand there is a version of TML Pascal that supports messages,
though it won't currently compile MacApp).  And there are translators
in the works that may allow using MacApp from other object-oriented
languages, such as Smalltalk.  But if you don't want to wait for
products that may never appear, MPW is the only choice.



Brian McElhinney
Software Guy
John Fluke Mfg. Co., Inc.
uw-beaver!fluke!mce
fluke!mce@uw-beaver

lsr@apple.UUCP (Larry Rosenstein) (01/21/87)

In article <320@escher.tc.fluke.COM> mce@tc.fluke.COM (Brian McElhinney) writes:
>Object Pascal and MacApp are sold separately from MPW, so the total
>cost is more like $300-400 (I didn't pay attention to the release

The current APDA prices are MPW: $100, MPW Pascal: $75, MPW C: $75, MacApp:
$50.  If you are doing development using MacApp, this makes the total price
$225.  These prices are going to be raised as of March 1, I believe.  (I
haven't gotten the latest APDA Log yet, so I don't know the exact new
prices.)  There is a free upgrade to the final version for people who
bought the beta releases.

>I have some problems with the lack other object-oriented extensions to
>Object Pascal.  The official line from Apple is that it confuses the
>poor programmer.  

The kinds of extensions that Object Pascal doesn't have are class objects &
methods, and multiple inheritance.  We specifically left these out of
Object Pascal in order to simplify the language.  There were 2 main reasons
for this.

First, if MacApp was written to use these "advanced" features, then
programmers learning MacApp would have to learn the additional language
features as well.  Class objects, metaclasses, and multiple inheritance are
known to be much more difficult to learn than the basic object-oriented
concepts.

The audience for MacApp and Object Pascal are experienced programmers, who
do not necessarily have any object-oriented background.  In addition, people
choose to use MacApp because it speeds up program development.  The more
learning time we add before people can use MacApp, the less attractive it
becomes. 

Second, it takes extra development time to add these features to the
language.  This is especially a problem for 3rd party language developers.
The more features we add to Object Pascal, the harder it is for them to
enhance their languages to work with MacApp.

In designing Object Pascal and MacApp we made some compromises in be the
"pure" object-oriented programming style.  Our primary goal was to build a
tool that Macintosh developers would find useful; learning time,
performance, and code size, therefore, were important considerations.

-- 
Larry Rosenstein

Object Specialist
Apple Computer

AppleLink: Rosenstein1
UUCP:  {sun, voder, nsc, mtxinu, dual}!apple!lsr
CSNET: lsr@Apple.CSNET

joel@gould9.UUCP (Joel West) (01/23/87)

In article <320@escher.tc.fluke.COM>, mce@tc.fluke.COM (Brian McElhinney) writes:
> If you want a copy, you should get your order in to ADPA, as they will
> likely be swamped (as usual).

Amen.  APDA has gone from 0-7,000 members, but Apple is going to pay
major penalties (particularly with its developers) if does not help
APDA solve its chronic logistical problems.

> The main advantage to MPW is that you
> can use Object Pascal and MacApp.  

The main advantages to MPW for most people have nothing to do with this:
	* It's complete
	* It's designed for big jobs
	* It's user extensible
	* The Pascal is 100% (not "99%" Lisa compatible)
	* The new tools are better than anything else available.
(It's also big, slow, and expensive.)

> Object Pascal is an extended Pascal
> (blessed by Wirth) with messages, but no other object-oriented
> extensions.  
> I have some problems with the lack other object-oriented extensions to
> Object Pascal.  The official line from Apple is that it confuses the
> poor programmer.  I feel that learning to program based on data
> objects that can respond to methods is what slows most programmers
> down, and obviously Object Pascal still requires that.  Once this is
> mastered, the more "advanced" concepts can be easily learned.  After
> all, aren't programmers supposed to be expert users?

Larry Tesler was quite adamant about this at the MacApp Developer's
Association meeting.  "We are not going to add something for 5%
the people. [programmers.]"  (perhaps a paraphrase).  One might not
agree with it, but I believe Apple's view is that they are producing
a journeyman product that is easy to use by programmers trained in
more traditional approaches.  As Tesler noted, object-oriented languages
have been around for 20 years without taking off.  And I think his
Byte article (8/86) suggests he sees himself as an evangelist (like
Alan Kay or Steve Jobs, in their own way) to popularize the technology,
because it holds enormous benefits if it would ever get used.

Besides, one advantage of this conservative approach is that it avoids 
what I call "kitchen sinkware."

Apple also has not prevented others from contributing.  While they may 
be the only ones to bring a supported Smalltalk out, other companies 
can solve other problems.  Although overpriced as a Macintosh development 
tool ($995), ExperCommon Lisp appears to be a complete object-oriented 
language along the lines of what Brian wants.

> Eventually, you may be able to program in Object Pascal without MPW (I
> understand there is a version of TML Pascal that supports messages,
> though it won't currently compile MacApp).  

TML 2.0 supports messages, but without conditional compilation, you'd
have to hand-edit the sources, and also would lose the MacApp debugger.
TML 3.0 might, but it's at least six months away.  Comments @ TML suggests
they might be more interested in being the only player in the IIgs market 
than the #3 (or #4, if Borland does well) in the Mac Pascal market.

> And there are translators
> in the works that may allow using MacApp from other object-oriented
> languages, such as Smalltalk.  But if you don't want to wait for
> products that may never appear, MPW is the only choice.

Apple is working on a "minimal C++", so I'd bet on it appear eventually,
but, I agree, just about everthing else is utter vaporware.

Apple, I'm sure, badly wants fileservers, and "desktop engineering", and
more and cheaper and better laser printers.  Third parties like Aldus
want to have the leading "desktop publishing" packages.  The comparative
importance of advanced development software is negligible.

Ironically, PPI has a product that might meet Brian's goals.  But despite 
being hyped in Kurt Schmucker's book and articles, Mac Objective-C is still 
vaporware.  Apparently some key PPI people still see the Mac as a toy.
-- 
	Joel West			     MCI Mail: 282-8879
	Western Software Technology, POB 2733, Vista, CA  92083
	{cbosgd, ihnp4, pyramid, sdcsvax, ucla-cs} !gould9!joel
	joel%gould9.uucp@NOSC.ARPA

chuq%plaid@Sun.COM (Chuq Von Rospach) (01/23/87)

In article <988@gould9.UUCP> joel@gould9.UUCP (Joel West) writes:
>In article <320@escher.tc.fluke.COM>, mce@tc.fluke.COM (Brian McElhinney) writes:
>> If you want a copy, you should get your order in to ADPA, as they will
>> likely be swamped (as usual).
>
>Amen.  APDA has gone from 0-7,000 members, but Apple is going to pay
>major penalties (particularly with its developers) if does not help
>APDA solve its chronic logistical problems.

Joel, you're a writer.  You should know better than to make that kind of
sweeping statement.  APDA is about six months old.  They are being swamped
by success.  But I certainly wouldn't call it chronic!  They aren't old
enough to have a chronic problem.  I expect that once the glut of
applications mellows out things will work out fine.  I turned in my
application the week APDA was announced, got my book early and have been
getting mailings on a reasonable schedule.

Now, what Apple SHOULD have done was create APDA two years ago, so the
pent-up demand for the organization wouldn't be quite so overwhelming.

APDA is a Good Thing.  Give them a chance to work it out -- they will.


chuq

Chuq Von Rospach	chuq@sun.COM

It's only a model...