We have a data acquisition system (DAS) interfaced with and controlled by a PC. The interface is via the parallel(printer)port. The data collection and analysis package mounted on the computer is now about 10 years old. It is completely reliable but runs under DOS on a PC with Windows 95 installed.
We are trying to replace the PC with a more modern laptop/notebook (this is essential for outdoor field work) but are having big problems in getting communication between the DAS and the PC. We do not expect the package to run under a modern version of Windows because after Windows 98 there is no true underlying DOS system. We have TRIED booting some newer laptops (several of them) from a MS-DOS boot disc and then running our program, but without success. The message from our PC is always the same “Cannot communicate with the DAS”. We have stripped the Windows system from the laptops and installed MS-DOS 6.22, but that does not work either. We have used the BIOS to configure the parallel port in different ways, SPP, EPP, Bi-directional etc to no avail. We are at a loss to know what to do next.
We would be grateful for any suggestions as to what the problem might be, and of course if anyone can suggest a solution then that would be superb. Revisiting the source code which was written in C and making all the necessary changes to the graphics routines etc so that it can be compiled to run under Windows is NOT an option.
So let me get this right: You want to make technology that by current computer aging standards is three generations obsolete, continue to work on new technology, with virtually no investment in design and coding? How long can this continue? Sorry to take a hard line, but perhaps your goal is fundamentally unrealistic?

All right, down to the attempts to get the thing running in the meantime.

You need this to run on a laptop, but presumably the laptop doesn't require much in the way of power. Have you tried an old laptop? Given the apparent requirements of your hardware, I would guess that an old 486 brick would do the job. You may need to try several before you find one that is sufficiently standards-compliant to pass all the necessary hardware messages through the OS. The good news is that you can pick them up for £25/$40 on eBay.

Alternatively, on newer systems, you need to try out the various DOS emulators. DOSBox should probably be your first port of call. After that, perhaps try FreeDOS or OpenDOS. OpenDOS would be more likely to work, but is harder to track down. The UNOFFICIAL OpenDOS site seems to have been ABANDONED quite recently.Yes, I suppose we are trying to make old technology(or old software) work in a 2006 environment, but we have been asked to supply a (final) system to an old friend and this is the last in the line so to speak. Hence we are unwilling to re-visit the source code and re-write the modules so that they will run under Windows. You have been very kind to reply at at all, and I am very grateful for that. What I do not understand is that if we take a modern laptop, strip the operating system off, re-format the disk and install DOS 6.22, we still have a problem. Is it a question of the SPEED of the parallel port or the CONFIGURATION of the port. or is it something entirely different All we want to do is to make this "last in the line " system work on a laptop/notebook computer that will allow experiments to take place in the field where there is no mains supply. We have made it work on a c.2000 desktop machine by getting into the BIOS and setting the parallel port to "bidirectional". However many modern laptops do not have a parallel port (they have USB only) and even in those that do have a parallel port, the options for setting the configuration are often limited.
We are expert in radio communications and "familiar " (but not expert ) with computing technology. This is why we are seeking help

DavidI guess my point is that you can probably make this work on some but not all laptops - not without re-coding. Laptop BIOSes vary MUCH MORE than PC BIOSes. PCs are dominated by no more than a handful of BIOS vendors, but almost every laptop has its own bespoke BIOS. And the APIs aren't generally published, so you're left with a lot of guess work.

Personally, I don't do any coding at the device level, and do not have detailed knowledge of this area. But to answer your question about the problem with the parallel port: no, I doubt that it is the speed of the port that is the issue. As far as I know, parallel port speeds have been maxed out for at least a decade (hence USB, Fireware, et al). No, the problem is more likely to be at the hardware interface level. A protocol conflict with your software, or an ignorant BIOS.

Try it on some older laptops though and I would guess you'll make at least one of them work. Once you know which one works, it's easy for you or your clients to pick up the laptops on (e.g.) eBay.We found an IBM Thinkpad with a Pentium 1 running at 100Meg and with Win98. Our software ran perfectly!!!
The thinkpad (like me) is past it best, but at least it works. The question now is whether our software is likely to run on a more modern Thinkpad running Win98 and with a, say, Pentium 2 at 300MHz. In other words is the CPU speed an issue? Presumably if we can get hold of a newer Thinkpad the BIOS will be the same, or similar, so the parallel port set up should be OK. It seems that we should not go more modern than Win 98, because after that we have no true underlying DOS system.

The bottom line however is that we have a solution; the question is can we find a better solution. Thanks for your help up to nowWell that's the $64m question.

Do you actually need Windows of any variety on the system? Have you tried just installing one of the DOS solutions? I'm thinking that you could set up bespoke systems that are dedicated just to running your software/hardware, SINCE no one is likely to want to use such an old laptop for anything else.

There isn't really any easy way to answer your question. I think you just have to experiment with some laptops. To be honest, if the P1 Thinkpad does the job, why not leave it at that? You can relatively easily come by them on eBay, and they are pretty cheap, so it should be easy to sell them as a solution for your clients. If your software works fine on FreeDOS/OpenDOS, you wouldn't even need to worry about O/S licences. Must be worth a shot.

Newer laptops will not successfully install Windows 98, due to the radical changes in the plug-and-play standards and power management standards (amongst other things). It would be impossible to say where the cutoff point is, because different manufacturers "upgraded" at different times.

It's a pain to recommend trial and error, I know, but since you who designed the system don't know exactly what hardware the system will work on, what chance do the rest of US have? Heh!

Good luck.