Re: In the Shallow End

Dan Johnson wrote:

"GreyCloud" <mist@xxxxxxxxxxx> wrote in message news:Q_qdnQwj8ZBja0LZnZ2dnUVZ_s2dnZ2d@xxxxxxxxxxxxxx

Dan Johnson wrote:

If you say so. I haven't read them. But I find your account
very vague; I still don't understand what about them made
them so good.

When a document claims how an API is supposed to be used and then gives the user examples that actually work, that is very good docs.
I've seen MS docs where a few APIs were documented but when used got the old not supported error message.
DEC docs were very accurate and never misleading.


You are still being quite vague.


It has been my experience a long time ago. Vague in your instance means you have no context to VMS or UNIX of that era.

I may say that I find that MS docs are quite accurate, and
they do specify which platforms each API is implemented
for. And their samples seem to work, such as I've tried.


MS docs are not accurate. Ever look at the DOCS in VS 6.0?
Terrible. Worse than their previous offerings... and their latest is a pile of words with no organization.
I checked out their latest VS C/C++ offering and their docs were in pretty sad shape as docs go.

[snip]

What simple functions are those?


Simple screen management functions, to name one.


Windows offers lots of this stuff. Indeed, it's one of its
main draws: it manages the graphics devices.

Before it was a nightmare and not complete.


It's got console support too, if you find the graphical
stuff to sophisticated.

Heh. Most of their APIs are way too clumsy to use in an efficient manner. Most programmers just have to put up with that junk. I feel sorry for those chaps.


DOS was kinda lame on this score, but MS's
products have got a lot better since then.


Not really, just more to go thru. How anybody can code with their tools is amazing to say the least.


When you have to roll your own because there isn't a viable product out there, the problems start to come up very quickly with MS development tools. I've had no problems with VMS and the excellent development tools.


What "viable product" is missing for you? There's an awful lot of
stuff out there for Windows, after all. And what 'problems' come
up very quickly with MS development tools?

At this current point in time, you can go out and buy anything you want in the generic sense of the word. But when you work for DOD, then you are in the realm of custom programming for a particular function. You won't find anything for sale out there. And again, MS tools are not all that good. The best stuff was written on VMS for JSTARS and also the Starwars project.


[snip]

Of course not. It is obsolete.

Not in VMS it isn't. Works great and is very easy to use.


It is obsolete on any platform. It *works*, sure,
but lots of obsolete stuff still works. The industry
has moved on from ISAMs.

In most instances, ISAM is very adequate for custom work.
If you want a decent database manager, you can get them for VMS from Oracle or Sybase. And even at that, they hook into the core file structure that makes their life so much easier.

Even MS used DEC equipment into the 90s before having to dump DEC when DEC sold out to Compaq.


[snip]

MS provides relational database tools instead; much nicer
that some old ISAM.

Guffaw!!
DEC did it the best. All MS could do was to make a lame attempt at a copy.


No. MS did not copy RMS because nobody wants that now;
people who proper relational database systems with SQL
support, transaction integrity, stored procedures, trigger,
constraints, the works.

They couldn't copy it because they know that they'd get nailed by DEC at that time or Compaq back in the 90s or HP today. It isn't one of those things they can incorporate.


So MS sells that.

[snip]

Could you be more specific about this? What conveniences
are you thinking of?

During program development. Even the DCL allowed for linkage of many different object modules together.


DCL was DEC's command language, and it beats
.BAT files hollow. But it isn't a linker;

LINK comes with the system.

you could
invoke a linker with it, but the object module
compatibility stuff was done by observing consistant
calling conventions.

That was done by DEC with all their languages. They were consistent across the board. You couldn't take a Fortran MS object code that used text and link with MS C object code that used text without going thru a few hoops. With DEC you could do that directly.


.NET, as I have mentioned, goes beyond this.


..NET is just a kluge to solve a few of their misperceived security problems.
And their solution still doesn't work.


Backups were easy.


Are Windows backups hard?

Yes, as the UI was not consistent or obvious.
I had to use the IOMEGA backup utilities instead with their tape drive.



Their All-In-One office package was better. etc.


I did not know they had one, still less that
it was better than Office is today.

A lot of people that did use all-in-one never wanted to leave it.
It was that well designed... just expensive as hell.


What was the name of this office package?
I'd like to look it up.

[snip]

Very probably inertia. VMS is a great OS, and I don't mean
to downplay it, but if they are still using RMS files for
they transtion engines, it's because they haven't bothered
to upgrade- if not to Windows, then at least to a relational
database.

This is where a lot of people don't get it. VMS was great in disaster recovery during 9/11. If it weren't for VMS, the stock market would have been irreparably damaged.


I doubt this very much.

There are Wall Street Journal articles around on the net for the looksee if you care to google for them.
And it was also posted in comp.os.vms.



As it was, a couple of days later all of the transactions were restored.


You can do this on Windows. This is
all well understood stuff; just a question
of shipping your transaction logs to offsite
replicates of your databases.

VMS networking was far superior and still is. Most of the IT pros know this and really don't want to leave that paradigm. If they do leave, they usually head toward Sun or IBM... but not to windows.


Though I doubt not that VMS was doing this
before SQL Server was!

The situation was fault tolerance and recovery. VMS excells at this.



That's why VMS isn't going to disappear. IBM is another strong vendor for this market and they do play a role in other market exchanges as well. Same concept, just different vendors.
What they like to do is use the Windows desktop as data gathering frontends to these operations.


Sure. Windows has great UI support, and you can run
all your favorites apps on the same computer you use
as the data gathering front end.


It was what the floor people were used to seeing so they implemented it. But as far as timing problems... well MS still has that problem. VMS does not.

[snip]

Because VMS is missing a lot of stuff MS
has; like the .NET stuff I just mentioned.

VMS is an operating system, not a programming language.
They don't need the .NET paradigm because their languages
are already robust, tried and proven.


By this I suppose you mean "old". :D


In this case there is no such thing as "old".
Just FUD.
C works well on VMS... and C works well on XP.
C++ works well on VMS... and C++ works well on XP.
Java works well on VMS... and Java works well on XP.
So what was your point again?

[snip]

Indeed not. Has some other language surpassed
FORTRAN while I was not looking? If so, which one?

Fortran is perfect for programs that can't afford to lose numerical accuracy. Otherwise your results will be perfect garbage.


You are quite mistaken about this. You'd actually
be better off with C#, and much better off with
COBOL, if you want maximum accuracy.

No, you are sadly mistaken. This is Fortrans forte.
C# won't check for bounds in complex arrays.
C# won't check for numerical accuracy thru large loops.
Neither will COBOL. COBOL isn't suited for scientific applications.
C# is not certified FIPS.


Fortran uses floating point binary numbers for
its data types, because they are fast. But their
accuracy problems are well known and serious.


Wrong again. Obviously, you aren't techically aware of what Fortran is.
Do you know what the IEEE floating point specs are?
DEC Fortran used several specs for certain instances... Like the G-Float, or H-Float. There is a lot more to fortran than you realize.
If Fortran is so dead, then why did Intel bother to produce the 2003 standard?

[snip]

That too. It was awful hard to bring that stuff to
the PC or the Mac. It was done, but it took a long,
long time. With Windows, you just drop in a card,
and install the driver for it. And it works.

MS wrote a very good basic interpreter for the Amiga. Best interpreter I've ever seen from them. Too bad they can't do a repeat now for the PC. I've yet to see multiple bit-plane animations done like the Amiga had done in its days.


Technology has moved on; but the revolution in 3d cards
that has changed PC graphics so much was only possible
because there was an OS to provide compatible access
to these cards.


Has nothing to do with the o/s. Just that MS has market dominance is all, and that is where their money is.

[snip]

I can't parse that. Certainly I agree that Commodore's
management had much to do with that companies fall.

Let me put it this way... Commodore spent too much money trying to produce a PC clone to run MS-DOS on in a saturated market. Before, they were making a decent profit.


That does seem strange and foolish; I suppose they
did this at the very end, when things looked hopeless?


Nope. Just bad management, as reported by the media.

[snip]

This is where they can't really innovate... all that luggage that prevents MS to push the technology further, and they have to carry it around. That's why Apple had to dump a whole paradigm to plunge ahead and take the lead. A very bold and risky move, but a successful one.


This is an interest allegation. What 'whole paradigm'
has Apple dumped?

The PPC paradigm... and the Motorola paradigm.
Kind of like the high risk that got Compaq in trouble when Carly committed Alpha-cide. One month the Alpha was a guaranteed thing, the next month it was the Itanium... better know at the time as the Itanic.

[snip]

The first Windows that supported timesharing with
a GUI for each user was NT 4, Terminal Services
Edition. That would appear to be the bit of mainframe
tech you are talking about.

It would seem so. But then, Bill had to *** in and push for moving the GDI into kernel space which made the o/s a bit unstable. It was then upto the graphics vendors to write good stable code for their grahpics cards.


No. The GDI was just as unstable when it was in CSRSS.exe;
an fault in that process caused a blue-screen, and the entire
OS halted. This is in no way an improvement over the kernel.


It was for speed. Back then it cost time on those cpus to change state priviledges. It is what got Cutler fired the first time. Later Gates realized how good Cutler was and had to hire him back on to complete the project of NT.

They were right to go with the kernel-GDI implementation;
it was a lot faster, and at the time that was critical.


Guffaw!! And now Allchin is getting it out of kernel space. Wish they'd make up their minds, eh?

[snip]

Yes. They also changed a few meanings as well. Back in the old days a subroutine was meant as such. A function was meant as such. But no... MS had to invent a new buzz word to muddy the waters... APIs. (Application Interfaces) I preferred the older meanings as everyone back then knew what you meant.


"API" does not mean "subroutine" or "function".

It's meaning is closer to "promise".

An API is an interface that apps use to interact with
the OS. MS promises to keep it stable so your apps
work.

Nope. Just a redefinition of old words.


It is composed of numerous functions, subroutines,
records, constants, and much else. But it is not any one
of its components.

[snip]

I think MS's long range planing was vidicated by events, really:
they did ship all that stuff, and they brought the industry with
them. It just took time.

But everyone was still waiting for MS to catch up where they had left off. Remember the old MP/M?


No, I don't.

But the customers do not, as a rule, like being rushed
from platform to platform. As a matter of fact, they hate
it passionately.


They do? Then we should just hold off Vista for another decade. Wouldn't want to stress out those people with such a drastic change.

[snip]

If you like. But that would have saved Java as a technology
from the dead end into which Sun was guiding it.

I doubt that Sun was guiding it into a dead end. Sun has always promoted Java, they just didn't want MS to derail their efforts.
By letting Sun manage their own language, and also develop it for many other platforms, everyone else can standardize on Java if they want to.
Can anyone standardize on any of MS languages now? They can't because of the reliance on their o/s which is proprietary.


Actually, they can standardized on C# if they want to.

No one can and it isn't advisable. If they did, MS would just come along and redefine the standard again.... this is called moving the goal posts.


(The list of volunteers is short, admittedly.)

[snip]

There is much in what you say, so I've snipped it. :D

But I'll object to this: Apple does *not* make computers
for the masses. They make computers for the elite.

The masses, as ever, want cheap computers and don't
care that much about quality, still less style.

That has changed.


I do not believe this.


That's your problem. I switched from windows to OS X because of the security and compact design of the iMac. Now I don't have to spend time or money on any vendors AV software.


Dells offerings aren't any cheaper than Apples offerings.


That's bad.. for Apple.


How so?

Think of the Caddilac Cimmaron. A rebadged
Chevy Cavalier which badly damaged Caddilac's
brand.


Car analogies aren't all that good.

If Apple is really selling cheap computers
that are just standard Intel parts with Apple's
logo slapped on, it's going to hurt Apple's
brand seriously.

Not the Mac Pro. It is priced well for the power, and Dell cost a tad more in that arena. So considering a $2.4k price tag I consider it a decent deal.


I do rather doubt that Apple will allow this
to happen, however. Apple will not follow
Dell as prices fall, I think.


As long as Apple gets their profits it won't matter.


I do see Apples in education systems more and more these days, so it looks like the ease of use is finally becoming obvious to the masses. It'll just take a few big box stores to realize this is all.


I think you are being overly optimistic here; Apple
has long had a strong presence in education, but
nothing much came of it outside the hallowed
halls of academe.

I wouldn't go that far... it goes with an old adage: give me children and they will stay catholic. Same goes for those taught in college or lower. Besides, that is what MS is doing as well. Why else would they give academics a price break?

--
Where are we going?
And why am I in this handbasket?
.