DOS/Windows and slashes (was Re: Windows XP Pro and PossiblefilePro Bug)

[Bill Vermillion]

On Sun, Dec 05 14:32  Fairlight said 'Who you talkin' to? You talkin'
to Fairlight?  I didn't do nuttin'.  I said: 

> At Sun, Dec 05, 2004 at 11:48:34AM -0500 or thereabouts, 
> suspect John Esak was observed uttering:

> > What makes the flavor of "convention" you like any more
> > important (or better design decisions) than those grown up
> > around a DOS/Windows environment???? When, in fact, those
> > -h or -- ones in particular are so much "newer" than the /?
> > which has been around since 1981 or 1982 as far as I can
> > tell. For you to put down something that arose long before
> > your preferred likes and dislikes as being something stupid
> > to be rankled over... is a bit parochial.

> I don't know why they didn't think to do something more
> sensible in the first place. CP/M had forward slash pathnames,
> didn't it? And that dates back to 1975.

I just checked a CP/M tutorial book to be sure I didn't
mis-remember.

CP/M can address 16 disk drives - A: thru P:

CP/M really is nothing more than a very crude file handler.

DIR A: lists the files on A, DIR P: list the files on P,
though I've never seen any CP/M system with more than 4 drives.

I will list ALL the comnmand available in the base CP/M

DIR, ERA, REN, SAVE, TYPE and USER.   

Programs normally supplied with CP/M are  ASM, DDT, DUMP, ED,
LOAD, MOVECPM, PIP, STAT SYSGEN, SUBMIT and XSUB.

So most of you OS file handling is done with the above commands and
writing in assembly.  Normaly CP/M distribution was on 
ONE single-sided single-density 8" floppy.   And that meant
256K max. 

All in all it was pretty crude, but people like Microsoft supplied
the languages so that you didn't have to learn assembly to do
anything worthwhile and of course dBase II was the big program
in use.

But there were no paths.

> Assuming my memory from my Turbo Pascal days (which required
> CP/M on the Apple) isn't too rusty, and CP/M did have forward
> slashes, then it makes little sense that MS should know about
> something that was a precursor to their OS and then flip it
> about so strangely. The only part that gives it any sensible
> context is the history of how it came about, as Bill put it
> forth yesterday.

Borland started it with about 3 different languages in the $50
range, one of which was Pascal. So if there were slashes in
the program they weren't part of the base OS.

> > Something that millions of users have come to use in "their"
> > environment, and you grouse about having to "do in Rome
> > as the Romans do" because you are used to something which
> > developed in "your" environment so many years later... well,
> > c'mon, you have better things to grouse about, don't you? :-)

> Mrm...yeah.  :)  And if not, I could always dig some up.  :)  

Well the - for options came before the MS slashes by about 10 years
- even before CP/M - it was from Bell Labs and it was called
Version 7 Unix.  [I don't know if those things existed in Version
6]

> Sometimes it's just the result of the oddity that is my
> computing timeline. I jumped straight from AppleDOS to BSD 4.3
> Tahoe, and -then- backwards to PC and MS DOS.

AppleDOS wasn't the best thing I'd ever seen.

> I know I worked on some TRS-DOS in there before *nix, but I
> don't remember terribly much of it.

The original TRS-DOS on the Model I was different from that on the
IIs.  The I started off with great ideas such as device
independance, and other Unixy things.  They are listed in the first
12 pages mimegraph pages I got with TRS-DOS 2.1 - a release that
was so early all the labels were handwritten.  About 6 weeks later
the first one available for sale had many things listed in the
first being nowhere to be found in the next one.

Randy Cook had great dreams, and his VTOS 3 was more like what he
wanted for TRS-DOS.  That begat VTOS 4, and the Lobo bought the
rights from Cook for LDOS - Lobo Disk Operation System - which
changed it's name to Logical Disk Operating System in a month or
so.

Here is something that will make you happy you live in today's
world of computer prices.

----
5 1/4" Winchester System
Compact and exceptionally reliable, with 4.8 megabytes of
high-speed Winchester storage plus a 720 kilobytes floppy drive.
The value leader in mass storage.
9501 for Model I or III with LX-80 (sold separately)  $3633.00
----
That's from the back page of the July 1982 LDOS quarterly.
In LDOS Quarterly July 1, 1981 is the announcment that LDOS
changed hands and became part of Logical Systems.

Oct 1981 issue advertises LDOS for the Model I and III
at $189.00.   That was only 2 months after the first IBM-PC came
out with PC-DOS at $65 and CP/M at $300.

But I can't find slashes anywhere.   Maybe they weren't invented
at that time. :-)  I'm beginning to feel that old.

> We spent most of our time in the BASIC interpreter, as it was
> a high school class--which was really amusing as I ended up
> helping teach part of it when we hit Pascal on the Apple IIcx,
> as the teacher was still taking his course on Pascal while
> teaching it, and I'd already known Pascal for three years. :)

I'd seen that approach before and locally the teacher had told me
he had heard the best way to learn a language was to teach it to
others.  That was JIT - Just In Time learning - but it meant that
any questions raised could not be answered by the teacher as he
wasn't there yet.   That seemed to be about the time things
change from 'education' to 'teaching' - the latter being the
'trade-school' approach of 'how to do something' while the former
has you undestand it all so you can tackle anything.

> At any rate, my migration path is odder than most people's, who
> went from the C64 or Apple right to PC first.

And you can say that because now most people have only been using
computers for 10 years or so.

> At any rate, my migration path is odder than most people's, who
> went from the C64 or Apple right to PC first. I more or less
> skipped that except for applications work like word processing
> and spreadsheets, and jumped back to the PC and MS DOS world
> after experiencing about four different flavours of *nix.

You still have time to correct that mistrake.

> Still...the "when in Rome" thing only goes so far.

I don't know about Rome, but a great many foreign countries
are moving to non-MS environments.

> There's nothing that says that a switch -must- start with a
> forward slash--except MS's utilities.

Microsoft was about the only place you could get a mix of languages
for the early PCs.  BASIC, Fortran, and Cobol come immediately to
mind, but the latter two surely seemed like an ill fit for
64K 8080 machines.

> People just want to broadly extrapolate and generalise rather
> than learn individual applications. COMMAND.COM doesn't
> actually care--it's the programs that care. I can use the old
> 16bit command shell with forward slash pathing with perl just
> fine, for example.

Well command.com doesn't do a lot other than call other programs. 
And forward slash and backward slash work equally well at the
'command' level in current MS offerings.  Now if they would
let you perform a CD with a drive letter all at the same time
we would be getting somewhere. :-). 

At only 50K the only .com file larger is edit.com.  There are only
16 .com files left in XP.   .com lasted far too long as it is.
Those show their CP/M heritage with their fixed starting points.
Command.com is more like a loader in many ways.

And - I just found another use for slashes.

I ran across a beta copy of DOSPLUS and it has a directory listing
attached.  All the extensions we are used to seeing in the PC
world with dots are slashes.  BOOT/SYS, BASIC/CMD
and I'd forgotten some of the bizarre directory listings.

An example
Filename   Attrib      LRL   #RECS  EOF DE FILE SIZE  MOD DATE TIME
--------------------------------------------------------------
BASIC/CMD  -IP---  EX  256      84  75  1  s=    21.0  17-NOV-83

I had forgotten about listings like that.  Information overload
describes it.

And I found a dual-format disk I had constucted.  I made the first
manually and the just did a pure dupe to use it.

I contstructed a CP/M directory entry and also an LDOS directory
entry and pointed them to the same start location on the disk.

Then I could just copy a file in CP/M and insert it in LDOS and
read it directly.  I'd forgotten how much I used to tinker
at the disk level.

That's how I learned how a lot of filePro worked.  In those days
with smaller machines and slower systems, I could watch things
happen.

On building an index you could watch it builda small file and sort
it and and write it - that's when we'd get those pauses building
and index - and then it would build another file in parallel, until
it had read all the records and built as many small files as
needed.  And then when it came time to write the new index it
looked at the smallest record on the top of each file/record 
and wrote it to the new file, and continued until it had written
them all.  FP was so much better about conserving disk space
than any other database manager in use at that time.

Bill

-- 
Bill Vermillion - bv @ wjv . com