DOS ain't dead

This one is for Bret Johnson.

I posted in other topic that I made Acrobat Reader for DOS run under vanilla DOSBox.

This is my runtime:
https://archive.org/details/acrodos-runtime

About the second topic: PDF 1.0 conversion, I'm also interested in that. But haven't found any software that's able to convert newer PDF formats to it or that is able ot genearte PDF 1.0. So, if you have any ideas, you can use this topic.

> I posted in other topic that I made Acrobat Reader for DOS run under
> vanilla DOSBox.
>

This time under DOSEMU2:

> I posted in other topic that I made Acrobat Reader for DOS run under
> vanilla DOSBox.
>

It also works under Qemu:

> I posted in other topic that I made Acrobat Reader for DOS run under
> vanilla DOSBox.
>

It also works in DOSBox-X

Just for the completeness sake, here it is in VirtualBox,too:

> About the second topic: PDF 1.0 conversion, I'm also interested in that.
> But haven't found any software that's able to convert newer PDF formats to
> it or that is able ot genearte PDF 1.0.

Exploring a history of Ghostscript for DOS, I found out I mostly have the binaries from FreeDOS repo. The first one of those to support PDF was 4.03. I tried 4.03 and 5.10, attempting something.

I tried converting modern PDF to PS with a current PDF2PS script. That didn't work as every generated PS featured some PS instructions these old Ghostscripts couldn't interpret.

A also tried converting a modern PDF to PS with the old versions of PDF2PS that came with those releases, but it couldn't really read them.

Anyway, I think these old versions might be able to generate PDF 1.0. The problem is, how to use this from the modern system? How to convert a well formed modern PDF file to use only the features of PDF 1.0 or how to convert PDF to a Postscript format that only include older versions of Postscript language?

This might not be the problem if you use an older program for DOS to generate PS and then try to convert it to PDF. I know some later word processor had support for generating PS (or printing to PS printers). Also some text editors have support for PS. I would guess there should also be some DTP programs or image editing programs.

But if you want to convert your favorite modern E-book to PDF 1.0, that, for now, is not possible.

> But if you want to convert your favorite modern E-book to PDF 1.0, that,
> for now, is not possible.

I actually had partial luck with this. You can actually convert modern PDF to layer 2 or layer 1 postscript with pdftops from poppler-utils or xpdf-utils.

pdftops -layer2 file1.pdf file1.ps

Than you can try converting generated postscript to PDF 1.1 with GhostScript 4.03 and its ps2pdf script. This will actually be able to be read in Acrobat Reader for MS-DOS.

That said, a lot of modern PDF files will fail to convert back to PDF 1.x. But some older documents might work.

> > About the second topic: PDF 1.0 conversion, I'm also interested in that.
> > But haven't found any software that's able to convert newer PDF formats
> to
> > it or that is able ot genearte PDF 1.0.
>
> Exploring a history of Ghostscript for DOS, I found out I mostly have the
> binaries from FreeDOS repo. The first one of those to support PDF was 4.03.
> I tried 4.03 and 5.10, attempting something.
>
> I tried converting modern PDF to PS with a current PDF2PS script. That
> didn't work as every generated PS featured some PS instructions these old
> Ghostscripts couldn't interpret.
>
> A also tried converting a modern PDF to PS with the old versions of PDF2PS
> that came with those releases, but it couldn't really read them.
>
> Anyway, I think these old versions might be able to generate PDF 1.0. The
> problem is, how to use this from the modern system? How to convert a well
> formed modern PDF file to use only the features of PDF 1.0 or how to
> convert PDF to a Postscript format that only include older versions of
> Postscript language?
>
> This might not be the problem if you use an older program for DOS to
> generate PS and then try to convert it to PDF. I know some later word
> processor had support for generating PS (or printing to PS printers). Also
> some text editors have support for PS. I would guess there should also be
> some DTP programs or image editing programs.
>
> But if you want to convert your favorite modern E-book to PDF 1.0, that,
> for now, is not possible.

This could be possible with windows. acrobat distiller 1.0 can parse output from *.prn file from "Print to file" function of "Apple LaserWriter" driver.
https://media.miniwa.moe/media/3a4adc6c1bf488731af70e685ab20aac8114c1a6209378798e02adfaf2763b66.png

> Anyway, I think these old versions might be able to generate PDF 1.0. The
> problem is, how to use this from the modern system? How to convert a well
> formed modern PDF file to use only the features of PDF 1.0 or how to
> convert PDF to a Postscript format that only include older versions of
> Postscript language?

I normally don't use PostScript, but I think the problem is that PS is really designed for printing and not viewing. I could be wrong, but I think you will lose the hyperlinks as you pass through PS. The hyperlinks are the only reason I'm even considering PDF (and maybe HTML) in the first place. If I don't have hyperlinks I'll just leave it as plain ASCII text.

I also find it interesting that, at least according to Adobe, all PDF readers are SUPPOSED to be able to read all versions of PDF documents. If an older reader is trying to read a newer document it should just be able to ignore what it doesn't understand, but that doesn't seem to happen (at least not from what I've seen).

Bocke:

Obviously your experiments with different VMs turned out differently than mine so the problem is not just with the VM itself. It could be a VM version, a configuration setting, the DOS version, the memory manager, or who-knows-what. I get errors and and/or the reader instantly exiting without even attempting to display anything.

> I normally don't use PostScript, but I think the problem is that PS is
> really designed for printing and not viewing. I could be wrong, but I
> think you will lose the hyperlinks as you pass through PS. The hyperlinks
> are the only reason I'm even considering PDF (and maybe HTML) in the first
> place. If I don't have hyperlinks I'll just leave it as plain ASCII text.
>

If you need hyperlinks, you'd probably have much more luck with HTML.

Converting from PS or PRN will kill all hyperlinks.

> I also find it interesting that, at least according to Adobe, all PDF
> readers are SUPPOSED to be able to read all versions of PDF documents. If
> an older reader is trying to read a newer document it should just be able
> to ignore what it doesn't understand, but that doesn't seem to happen (at
> least not from what I've seen).
>

That's interesting, but it doesn't work in practice. Older versions of Acrobat Reader will crash if you try to open a version of PDF they do not support. It's not specific to DOS version.

> Obviously your experiments with different VMs turned out differently than
> mine so the problem is not just with the VM itself. It could be a VM
> version, a configuration setting, the DOS version, the memory manager, or
> who-knows-what. I get errors and and/or the reader instantly exiting
> without even attempting to display anything.

True. But at least in DOSBox, there is only the internal version of DOS and internal memory management. Unless you are using an old version of DOSBox or your host doesn't have the drivers for your GPU, it should work the same.

> This could be possible with windows. acrobat distiller 1.0 can parse output
> from *.prn file from "Print to file" function of "Apple LaserWriter"
> driver.
> https://media.miniwa.moe/media/3a4adc6c1bf488731af70e685ab20aac8114c1a6209378798e02adfaf2763b66.png

If the otput is a PostScript file, it could also be converted to PDF with ps2pdf that comes with GhostScript.

> True. But at least in DOSBox, there is only the internal version of DOS and
> internal memory management. Unless you are using an old version of DOSBox
> or your host doesn't have the drivers for your GPU, it should work the
> same.

I agree that's how it SHOULD work. I'm using DOSBox 0.74-3 and when I start the Acrobat DOS Reader the entire DOSBox window crashes and shuts down. In DOSBox-X 2022.08.0 I get an "JMP illegal descriptor type B" error and it crashes and shuts down.

I'm pretty sure I'm not using the latest versions of those, and if I upgraded (or downgraded) either of those the problem might go away.

> I agree that's how it SHOULD work. I'm using DOSBox 0.74-3 and when I
> start the Acrobat DOS Reader the entire DOSBox window crashes and shuts
> down. In DOSBox-X 2022.08.0 I get an "JMP illegal descriptor type B" error
> and it crashes and shuts down.
>
> I'm pretty sure I'm not using the latest versions of those, and if I
> upgraded (or downgraded) either of those the problem might go away.

No, the versions are ok. I use 0.74-3 too and it works.

It might be something connected to drivers or programs you load in DOSBox. Try using it bare with empty [autoexec] and [config] sections in dosbox.conf/dosbox-x.conf and without autoexec.bat.

What OS?

I am sorry if I'm bothering you with this. I just think there shouldn't be that big difference between similar versions of DOSBox and that it might be easily solvable. Maybe I'm wrong, but my tech support senses are tingling. If you don't want to bother, it's ok. I won't ask you anything about that anymore. :)

> > This could be possible with windows. acrobat distiller 1.0 can parse
> output
> > from *.prn file from "Print to file" function of "Apple LaserWriter"
> > driver.
> >
> https://media.miniwa.moe/media/3a4adc6c1bf488731af70e685ab20aac8114c1a6209378798e02adfaf2763b66.png
>
> If the otput is a PostScript file, it could also be converted to PDF with
> ps2pdf that comes with GhostScript.

the problem is GhostScript doesn't generate PDF-1.0 files.

Hi:

I suppose you know there is a DOS port of MUPDF:

https://code.google.com/archive/p/nanox-microwindows-nxlib-fltk-for-dos/downloads?page=2

It can open more modern PDFs than only 1.0 version.

Regards.

>
> I posted in other topic that I made Acrobat Reader for DOS run under
> vanilla DOSBox.
>
> This is my runtime:
> https://archive.org/details/acrodos-runtime
>
> About the second topic: PDF 1.0 conversion, I'm also interested in that.
> But haven't found any software that's able to convert newer PDF formats to
> it or that is able ot genearte PDF 1.0. So, if you have any ideas, you can
> use this topic.

Program Halibut can perform conversions to PDF format. The DOS version also exist.

---
DOS-u-akbar!

> I suppose you know there is a DOS port of MUPDF:
>
> https://code.google.com/archive/p/nanox-microwindows-nxlib-fltk-for-dos/downloads?page=2
>
> It can open more modern PDFs than only 1.0 version.

I know there are other options, but I'm trying to figure out a way to do what Adobe originally intended and said should work. It looks like it may not be possible, though.

> No, the versions are ok. I use 0.74-3 too and it works.
>
> It might be something connected to drivers or programs you load in DOSBox.
> Try using it bare with empty [autoexec] and [config] sections in
> dosbox.conf/dosbox-x.conf and without autoexec.bat.
>
> What OS?
>
> I am sorry if I'm bothering you with this. I just think there shouldn't
> that big difference between similar versions of DOSBox and that it might be
> easily solvable. Maybe I'm wrong, but my tech support senses are tingling.
> If you don't want to bother, it's ok. I won't ask you anything about that
> anymore. :)

I did some experimentation and figured out the problem. It turns out to be the DPMS servers I use (which were made by Novell back when they were in charge of DR-DOS). Apparently, DPMS has a conflict with whatever DOS extender Acrobat Reader uses (I saw somewhere what it was, but I don't remember - but I know it's an extender and not DPMI). I don't know if you're familiar with DPMS or not, but it's sort of a "DPMI for TSRs and device drivers". I'm converting all my TSR's to take advantage of DPMS since it can save a TON of memory.

> I did some experimentation and figured out the problem. It turns out to be
> the DPMS servers I use (which were made by Novell back when they were in
> charge of DR-DOS). Apparently, DPMS has a conflict with whatever DOS
> extender Acrobat Reader uses (I saw somewhere what it was, but I don't
> remember - but I know it's an extender and not DPMI).

It is called DOS/4G from Rational Systems (later Tenberry Software).

Quoting: In case of problems, DOS/4G or DOS/4GW can be replaced with the newer and free DOS/32; a patch utility can even replace DOS/4G code embedded inside a compiled executable file.

A DOS/32 fork can be found at https://github.com/yetmorecode/dos32a-ng

---
Forum admin

> the problem is GhostScript doesn't generate PDF-1.0 files.

Not sure. The first version of GS to support PDF was one of 3.x releases, but I can't find the source or binaries for that. GS 4.03 by default generates PDF-1.1. I actually managed to open one of those generated files in Acrobat Reader for DOS. It might depend on how complex the source PostScript file is though.

> I did some experimentation and figured out the problem. It turns out to be
> the DPMS servers I use (which were made by Novell back when they were in
> charge of DR-DOS). Apparently, DPMS has a conflict with whatever DOS
> extender Acrobat Reader uses (I saw somewhere what it was, but I don't
> remember - but I know it's an extender and not DPMI). I don't know if
> you're familiar with DPMS or not, but it's sort of a "DPMI for TSRs and
> device drivers". I'm converting all my TSR's to take advantage of DPMS
> since it can save a TON of memory.

I had the best luck with leaving memory management to DOSBox. It usually doesn't like 3rd party memory managers.

Edit: A small correction of the text.

Note that my documentation experience is almost nonexistent, so I do not claim to be an expert.

> I normally don't use PostScript, but I think the problem is that PS is
> really designed for printing and not viewing. I could be wrong, but I
> think you will lose the hyperlinks as you pass through PS. The hyperlinks
> are the only reason I'm even considering PDF (and maybe HTML) in the first
> place. If I don't have hyperlinks I'll just leave it as plain ASCII text.

Isn't PDF a superset / replacement for PostScript (PS)? It has a LOT more features: bundled fonts, compression, scripting, etc. Both are meant to be shown or printed exactly (and thus unlike HTML which is very loosely rendered). It's way overkill, IMHO.

Emacs used to use Texinfo, but apparently they thought about changing a few years back to AsciiDoc. I have no idea what came from that discussion. (DJGPP did just recently have yet another GNU Emacs binary published.) Emacs' Org Mode is also a good alternative, apparently.

Rob pointed to Georg's port of MuPDF, but there's also his FlWriter which supports a few documentation formats.

Another good option (that sounds like exactly what you want) might be pdfTeX, which is apparently bundled in most modern LaTeX distributions (using Knuth's TeX behind the scenes).

> I had the best luck with leaving memory management to DOSBox. It usually
> doesn't like 3rd party memory managers.

DPMS is not really a third-party memory manager. It is more of a "Protected Mode Application Environment", not too different than what DPMI and DS Extenders provide. E.g., you can directly use the functions provided by standard memory managers (like EMS & XMS), which is what a lot of older programs did. Or you can use a Protected Mode "Environment" like DPMI or a DOS Extender which "leverages" what the memory manager provides to make your life easier. The "DPMS" environment is specifically designed for the special requirements of TSRs and Device Drivers, allowing the computer to switch in and out of protected mode when the TSR needs to do things in the background. Like I said, DPMS is sort of like "DPMI for TSRs". It doesn't try to bypass or displace memory management.