Is there a need for new versions of PB/DOS?

[Michael Torrie]

I'm writing this to question the need for further development of PB/DOS. Bear in mind that I'm not questioning the need to continue to develop in PB/DOS. PB/DOS-based programs are continuing to be written to fulfil needs, many of which are for special-purpose applications. However, does this existing user base really justify a further version of PB/DOS? From what I can tell, based on my knowledge of computer languages in general (and the fact that all languages can be boiled down to two operations, which operations I forget), PB/DOS is a very mature and full-featured language. Except for OOP stuff, the language is complete. There are many syntac sugar things that could be added, I suppose, but it seems fine as it stands now.


Judging by the amount of traffic on this forum as compared to a year ago, or on the other forums, it seems like fewer and fewer people are developing in PB/DOS. And those that are seem to be working in DOS boxes within Windows or OS/2. Except for the OS/2 and Linux people, perhaps most developers should ultimately move to PB/CC or PB/DLL. DOS is dying, at least as far as most newer users are concerned. I realize the upgrade will cost something. For us hackers, DOS will be around as long as need be in some form. For that we can continue to use our existing version of PB/DOS, say version 3.5 or 3.2. While I once longed for a better editor, 3rd party programs have now emerged and function as well as I would ever want. Therefore I will continue to use PB/DOS (albeit now in a DOS emulator on Linux) but I have no plans to ever upgrade (and I'm still on version 3.2). I would rather have a native PB compiler for my platform (if I used windows, I'd be switching to PB/CC probably). (I know Linux is in the works, so I'm not asking about anything relating to it.)


It seems to me that PB should focus all it's energies on the current platforms, of which DOS is no longer one of them. The existing user base does need to be supported, and this is what this forum is about.


Had PB decided to add, say object-oriented extensions to the language, that might have been grounds for a whole new version, had it been last year or the year before. (If there is a new version with such features, I'll look at it I think the window (no pun intended) of DOS upgrades is past.

Michael


------------------


[This message has been edited by Michael Torrie (edited September 11, 2000).]

[Mike Luther]

A different view of why PB 3.5 needs to be thought forwarded in its
present format ..

Two weeks ago I was handed an edict by one of my major medical site
users. They politely told me that although they intended to go forward
into the future with hospital-wide picture integration and so on, they
were, it seems Federally told something. Both I and they will have to
support the current DOS based system - either native or embedded into
Novell, WIN, OS/2, whatever .. for a minimum of SEVEN years from the
date they make the change.

The shoe is beginning to drop on the position that irrespective the
data was stored, for medical records purposes, it has to remain in
state, fully accessible, for a statutory period of at least seven
years as-is etc..

Now I wanna move forward too, and I am .. However, I have to maintain
backwards level compatibility with the whole system for at least seven
years for every change I make.

So much for getting rid of DOS and the need for PB 3.5, since most
of this stuff doesn't run or isn't intended, in reality for WIN based
systems. If anything .. UNIX (LINUX) is, it would appear, a perhaps
better choice for upwards in the busimess than WIN .. but .. to keep
from running afoul of the Federal regulators, the NARA act, and so on
in the USA, good old DOS will be alive for a long time yet in these
system, or else. It's still a common focal point, even though all
are trying mightily to kill it.

Just another viewpoint .. strictly personal, from someone who will have
to comply or else.


------------------
Mike Luther
[email protected]

[Gary Akins]

Also - DOS may be fading on the desktop scene... but I assure you, DOS still holds a big presence in the embedded-systems world. There are all kinds of devices which use x86-based microcontroller systems (from 8088's to 486's) to perform a variety of consumer, scientific, and industrial applications, and DOS still holds considerable sway in that world. (Although Linux is starting to attract attention now that some companies are offering "Hard" Real-Time extensions to it.) Most people just don't realize it, because devices like these don't look like computers - typically, they're "black boxes" without full-sized keyboards, monitors, and other such accoutrements - and you never really see what's under the hood.

------------------

[Michael Torrie]

I understand that DOS does have a huge presence in the embedded word, and for situtation such as Mike Luther is in. Again, I wasn't questioning the need to maintain or develop software for DOS using PB/DOS. My only point is that DOS hasn't changed in many years (no new features to the OS) and that PB is already a complete and mature language. Any real new additions to the language are merely syntactic sugar at this point. Therefore I don't believe that any new versions of PB/DOS will really be needed. Embedded software will continue to flourish in the DOS world using the existing tools. The rest of us (it seems) have and should migrate on to other platforms. (It pains me to think that this implies moving to Windows -- I'm a DOS hacker from way back)

On another track, one of the reasons I've been programming in PB/DOS for so long is that I grew up on DOS, using text-mode based stuff with simple but powerful APIs. When everyone started moving to windows, the complexity of the windows API and the difficulty in writing even a Hello World program kept me from doing any windows programming. So I was excited to discover Linux, where my old DOS skills still applied, and using clean APIs, I could write everything from simple Hello World to full-blown GUIs without the messy win api stuff. I learned there were actually clean tool kits out there for making greate gui apps, and porting back to windows was a lot easier. Of course, now with PB/CC, many of us DOS guys can now program with minimal learning curve on windows. Except now I'm not using windows anymore.

Anyway, good luck to the PB team and may they design good products for many platforms, even maybe DOS.

------------------