Legacy DOS program capability preservation

[epineh]

Have a look at the TurboCNC manual, scroll down to the end, it tells you how to create a dual boot Win98/Dos selection at startup. I didn't post it here as it is a little long winded, which is why I just use F8 at startup and choose.

Russell.

[rancherbill]

Here is the technique to get the drive size that you need.

http://www.seagate.com/support/kb/di...8_4/index.html

This will solve the last of your remaning issues. I sure othe drive manufacturers have similar software for use with their drives

I re-scanned the total thread and everything that you need has been posted. There are several different good approaches and they all will work.

The only thing left is for you to pick one approach and get on with it.



Daedalus

The DOS software plays with the hardware. Virtual PC, which is a neat concept, won't work in this case.

[ger21]

Originally Posted by NC Cams Dr Motion: I plan to save the data but d/l'ing to disks is just too impractical. Besides, some files now can't be saved to disk unless you span disk and I've had some issues with that on occasion. Once burned, forever shy. If I could burn a CDROM via DOS and/or Win 95 (neither is capable of burning a CDROM), that would suit me fine as I could d/l the data, FDISK, revitalize, whatever. HOWEVER, I can't backup my critical data!!!!

I believe he said to plug your failing drives into a new PC and copy the files hard drive to hard drive. Once copird, put them back in the old PC's.

[fyffe555]

Hi NC, sounds like there's some confusion on what's what. As a sad old sod who thought DOS 5 was a huge leap forward I can add some info, or add to the confusion. I'll let you decide which!

When you say you have a DOS app that requires to run on DOS with a HD supported on the MB and DOS version you have, you've got something that uses ISA as an input/output with hardcoded card support? Similarly there could be some custom i/o? (disk writes/reads directly from the application to the drive?) That was a fairly common practice way back when.

You have a couple of limitations with drives;

DOS came with fdisk, the program which could create a partition on the drive which i know you;re familiar with. The maximum disk size depends on Fdisk version however and NOT DOS OS version. Early Fdisk versions with dos suppported 525mb, then 2.1, 4.2 and 8.4gb in 6.22, but partitions of only 2 or 4gb depending on version.... again note thats the Fdisk version and NOT a DOS limitation. Its all about how the hardware addresses the drive and how and what boot MBR, FAT or partition tables are written. Dos will read a fat 12, 16 table just fine. You can add fat32 support. You can use an NT or win98 version boot diskette and partition the drive to Fat32 8.4gb and dos 6.22 will read it fine depending on version if the hardware and fdisk will present the drive in an acceptable format.

The harddrive size is first supported by the DOS-bios on the MB. Note thats NOT the DOS OS. DOS-bios is now just BIOS and loads the mb on startup. The hard disk drive must be supported by the computer's ROM BIOS APIs, which have a 1024-cylinder limitation, in order for FDISK to partition the hard disk. If your MB is an early board and bios & IDE adapters doesn't support LBA then you're stuck with the IDE onboard limited to a max of 8.4 as I think you've already discovered, or get the LBA drive to appear as non LBA - 1024cyl max. A modern high capacity LBA IDE drive can still be partitioned into up to 24 seperate partitions that dos will recognise if it can be presented as a non LBA drive.

You can in theory upgrade bios to support LBA and I'm assuming you've tried that and not found a bios to suite. I know I couldn't when I tried the same thing.

You can put fat32 support into dos 6.22 along with long filename support and
address modern disks - as long as the hardware ide bios spi will support it.

You can dump the onboard MB bios ide limitation entirely and get a third party card IDE adaptercard like an older promise Ultra33 or 66 or Acard ata66. These cards load thier own bios after the MB bios loads and the cards do natively support LBA drives up to 128mb. Your partition limit then becomes what dos will address. Drives can coexist on MB ide and card ide and you can set boot order. FDisk will recognise drives on card natively without drivers as long as the card bios loads. You can define partitions on any new parallel ide drives and dos will recognise them. Any custom i/o to partition will work too. Problem is the cards are PCI so do you have a pci slot as well as ISA? Further Warning; I have this working on a P166 running dos 6.22 & fat32 and TCNC on an 8.4 partition on a new 160gb drive. It took some effort to get the sequence for attachment and fdisk operations for it to work but once done its stable.

You could swap to a newer MB with isa, pci and lba ide support for new drives. A lot of the later PIII slot1 mb's had that . I'm running exactly that on an abit mb for a server.

Newer drives are often available with a bios routine thats loaded on the drive to appear as an older non LBA drive - 1024cyls. Maxtor. seagate and WD have this available as part of the retail package and its downloadable from thier support sites. It does work but I've had inconsistant results with it.

To backup I'd be very careful. Any attempt to load new software on an older failing drive could possibly damage your data as I'm sure you know. Any attempt to remove stuff or add drives to a old established working machine can cause it to not play anymore. So any attempt to add new tape drive, DOS cd rom burning or what ever would be inadvisable. I'd go with Ger's suggestion to carefully remove the drive and add it into another machine and take a backup on a different machine. Safest way, and assume you've got to start from scratch and 'rebuild' the thing again.

There are some free dos cd burner programs but they come and go and I don't have a link to one right now.

I would NOT suggest using disk caddies for ide drives for backup as they DO blow drives unless the power is entirely off the machine, that is not just switched off but unplugged from the wall or whatever. they are not hot plug or even warm plug.

If you've got real i/o to disk or isa card you cannot run your app under xp, win98, VMware, linux emulation or whatever...

CF / Compact flash to ide adapters ONLY work if they are used as a read only device or for very limited writes in a pc. The write cycle life is about 200k. Win98 could do that number of writes in a weeks use...... The IDE mapping wouldn't necessarily map id custom i/o to ide is coded in the app.

Finally I use some of these; http://www.memorydepot.com/details.asp?id=DOM1GV40. Expensive, yes. reliable. yes. Have some running as boot disk on a communications server for four years. Write cycle time was rated as 3m, I 've exceeded that many times over and they still test clean. Totally compatible with older DOS/ IDE 40 pin interfaces. They might be a band aid but a better option than 'refurbished' drives...

Andrew

[fyffe555]

Had a thought this am. You can network DOS machines quite easily, it was done. Look for Lan Manager from MS for ethernet cards or IBM's Lan Server for token ring. Once installed it will allow a one way or two way 'share' of one disk from another machine. Add a quick .bat routine you can automate the movement of copys for backup... I might still have lan manager somewhere back home...

[NC Cams]

Fyffe: there are several machines involved but they all run the same software.

The cam measuring machine uses a special Heidenhain card to read linear and rotary encoder signals to "read the cam". This card runs in an ISA slot. It is hardlock linked to/with software via either parallel port dongle or an ISA card that plugs in.

The guy who wrote the software learned the hard way that XP wouldn't work with the original cam measuring hardware and/or the software because of the (lack of) support that XP provided via parallel port as opposed to what DOS provided - recall that he was using a parallel port dongle.

Besides, the cam design and analysis software does some fancy memory swapping - the program was developed BEFORE they had HIMEM managers and the program categorically won't run/load high.

So, no matter how much memory you have, you're only using the 640K limit - thus, you have to be REAL careful when you start loading/using other software and/or especially Win 3 programs (where my dynamics analysis programs reside).

It is real easy to get NOTHING to work because you run out of memory or something uses someone else's memory alottment. After many "what the frock??" moments, we simply built "dedicated" DOS machines and left well enough alone.

After weeks of trying and numerous WTF's over the years, we learned that ANYTHING BUT DOS 6.22 in concert with a carefully configured Win 3.11 system will cause you unfathomable grief.

This software is so sensitive (dare I say primitive) that we've even had M/B's with certain chip sets not work with it. It is quite simply easier to find what works and use ONLY that configuration. Better that than to mess with interrupts, memory mapping and goodness knows that else we've screwed to get the programs to run in a stable fashion.

Which is why we want/need/will only make a dedicated, carefully configured DOS machine with Win 3.11 for this particular system. This is also why we seriously try to avoid add-on cards and other gingerbread - it is too easy to add something and then have NOTHING work for no viable reason.

BTW, a lan card was tried and the memory it allocated to/for its use created havoc with the cam software so we had to abandon it.

This was why we had hoped to be able to at least get a CDROM burner to work somehow - simply burning data to disks for B/U, as much of a PITA and or expense that it can be, is easier/safer to do than to deal with the alternative created by dated, unforgivingly written DOS software.

BTW: I've recently learned that I can get "updated" cam design and analysis software. HOWEVER, I'll need to buy new Heidenhain PCI based cards, new software and all new PC's and convert to XP - to convert to "contemporary" XP based software that will only set me back close to $15K.

An expenditure that I simply can't justify spending in order to do NOTHING technially better than what I can do with lame a$$ yesteryear DOS software.

But I would have MARVELOUS graphics - for a simple, albeit static graph on a screen, hardly something I can justify spending that kind of $$$'s on for so little merit.

[lerman]

Hmmm. Perhaps this is a summer project for a college student.

Get a board to interface to the disk port on the ISA controller. Mesa makes a generic FPGA board that should be fine. That's a PCI board that you can put into another machine (running linux, or XP, or whatever). The board is about $200 in qty one.

Then write an IDEdisk emulator that gets the commands from the IDE channel and converts them to reads and writes on the second system.

Unless the DOS system is very sensitive to timing, this should work just fine.

It would actually be a pretty neat little project. I'll be you could sell a few of them.

(I know, this isn't the type of answer you are looking for. But if you have a friend at some local college, it should be doable as a summer project for some student.)

Ken

[NC Cams]

WIth all due respect and with tongue firmly in cheek, when someone says:

"...Then write an IDEdisk emulator that gets the commands from the IDE channel and converts them to reads and writes on the second system...."

My eyes glaze over. I'm not into programming which is why I got into making camshafts. The computer programs that I use came about as a GUI based version of a technogeek/technobabbly PC adaptation of what used to take a frigging mainframe to calculate.

I see the same look developing in peoples faces when I start taking 5th order polynomial exponent coefficients and ramp junction factors when I start talking cam design stuff. Not only do the eyes glaze over but they literally roll back into the client's heads.

My clients want a camshaft made that will make more power and not tear up vavletrains. It can be simple to create something like that when you know how - if not, guys create garbage and broken pieces with a trail of carnage. I simply create techno-garbage when I try to program a PC - classic case of "right guy for right job".

I have a legacy system. It works. It works the same DOS 5 as it does in the latest XP varuabt BUT at far less cost and/or complexity.

Sadly, it doesn't need anything more complicated than ISA and 4-6 gig of HDD to run. Hell, it will run as well and/or as accurately on a 486 as well as on a Pentium 333. All I need is some new "old" HDD's and I'll be good for, at the rate things are going, the rest of my career.

The ONLY reason why computers got to be SO popular and SO pervasive in our society is that Jobs, Gates et al figured out how to make them useable WITHOUT having to have a college degree. Point and click, plug and pray. For that, they are to be commended.

For their inability to make it simple and needing of a computer science PHD to understand how to program the damn things, they need to be shot. How could someone be SO brilliant in one stretch of the adaptation scenario and so irresponsible and/or undisciplined in others???

Simply too much time is spent innovating to obsolete today's parts tomorrow and not enough thought to maintaining interoperability with legacy systems over time.

I contend that some legacy stuff is quite adequate even though it may contain yesteryear technology.

Imagine what would have happened to our phone, TV and AM/FM radio systems if the prior systems had gone obsolete every 3-5 years or so???

Or what would happen to the cost of air travel if 3 year old planes couldn't fly anymore because the avionics was totally obsoleted (it is but there are ways and folks who still service it, thankfully or else the Defense Dept budget would be even more astronomically bloated than it is).

Post #50 offered some accute insight into the true sense of the effort so far. Via public and private e-mails, I"ve found the answer to my problem thanks to the plethora of knowlege contained in the minds and shared by a lot of well educated and experienced members of this community.

Thanks to all those who've contributed.

Surely, someone else may/will have a similar need some day.

The following links that got contributed are of specific value to those of us stuck in DOS dome:

http://support.microsoft.com/kb/2172...131&sid=global

http://support.microsoft.com/kb/255867

These helped a bunch. I think I have a line on the hard and software that I need to carry on with my lame ass yesteryear system in a manner that should last for quite some time to come.

Trust me, I learned my lesson about long term DOS system survivalism and I"m making a lifetime buy of my next system "fix".

[Willbird]

I know some of the same issues you talk of existed at usgov.org on military installations, because computers and hard drives can with great difficulty go IN, but they never come OUT in operating order for opsec reasons. So they did stuff like adapt 486 cpu's to 386 computers, and other weird stuff. We got some units that were scrap out of wright patterson AFB, the cases were built literally strong enough to park a car on, the power switches were all hot glued in the ON position..the cases were fastened shut with anti-tamper screws, on and on....interesting hardware tho...mostly 386, some had add on ram cards that held 2 megs of ram and each card went into the ISA bus, and the card was as long as an AT mainboard, not a half sized one :-).

there was a flurry of Cobal and other language hiring around y2k to deal with that issue on legacy DOS stuff like you have NC, and for much the same reasons.

Bill