Printers talked postscript. Now I hear they talk PDF. I know that it is trivial to convert from ps to pdf. Do programs that want to print
generate the PDF directly, or do they generate the ps and then convert
it to pdf?
It also occurs to me to ask what is the advantage for printers of
talking pdf instead of ps.
[...] I know that it is
trivial to convert from ps to pdf. [...]
It also occurs to me to ask what is the advantage for printers of
talking pdf instead of ps.
Thanks :-)
On 10/09/2026 20:57, Carlos E.R. wrote:
[...] I know that it is
trivial to convert from ps to pdf. [...]
To my understanding, it's actually the other way round. PDF is, or at
least was, a wrapper format around a subset of postscript. To get
equivalent postscript, you just need to unwrap the postscript, and give
it the extra stuff the pdf contains (like images or fonts) in a way it understands. To convert arbitrary postscript into a PDF, you need to
rewrite the postscript to only use the non Turing-complete subset
allowed in pdf, which I could see running into the halting problem and
other fun stuff. However, once you do have the final result of the postscript execution, the pdf wrapping is indeed quite simple, with the
most complicated bit being the cross-reference index that is "just"
counting how many bytes away each object is from the beginning of the file.
It also occurs to me to ask what is the advantage for printers of
talking pdf instead of ps.
The big one I know is that PDF is not Turing complete, while PS is. You don't want your printer running infinite loops or memory bombs or
similar. A malicious postscript document doesn't need a buggy postscript interpreter to do it's dirty work, a malicious pdf needs to cause the interpreter to misbehave to do any damage.
Carlos E.R. <robin_listas@es.invalid> wrote:
Printers talked postscript. Now I hear they talk PDF. I know that it is
trivial to convert from ps to pdf. Do programs that want to print
generate the PDF directly, or do they generate the ps and then convert
it to pdf?
Neither. Programs shouldn't care what format the printer can handle.
That's up to the cups printer driver, to take in what the program
generates and convert it to what the printer can deal with.
But I suspect any printer that can handle PDF can also handle postscript
(and plain text and probably PCL) so it's not "PDF or PS" but "PDF and
PS".
Heh, the HP LJ MFP-4301fdw claims "HP PCL 6, HP PCL 5e, HP PCL 5c, HP Postscript level 3 emulation, PDF, URF, PWG Raster, Native Office"
It also occurs to me to ask what is the advantage for printers of
talking pdf instead of ps.
I suspect more for printing from mobile devices; PDFs are common (eg
bank statements) so someone who is "phone first" may download a PDF
document on their phone, view it, print wirelessly. But even with
firefox on Linux, printing may generate a PDF (I just disabled the
queue, did a test print, saw it was a PDF in the queue).
To convert arbitrary postscript into a PDF, you need to rewrite the postscript to only use the non Turing-complete subset allowed in
pdf, which I could see running into the halting problem and other
fun stuff.
The big one I know is that PDF is not Turing complete, while PS is.
You don't want your printer running infinite loops or memory bombs
or similar. A malicious postscript document doesn't need a buggy
postscript interpreter to do it's dirty work, a malicious pdf needs
to cause the interpreter to misbehave to do any damage.
On Fri, 11 Sep 2026 15:24:23 +0300, magardner2010 wrote:
To convert arbitrary postscript into a PDF, you need to rewrite the
postscript to only use the non Turing-complete subset allowed in
pdf, which I could see running into the halting problem and other
fun stuff.
Adobe had a product called ?Distiller? which did this. You ran a
custom wrapper program on a PostScript interpreter, which I think
intercepted all the actual rendering calls and output corresponding
PDF objects instead.
The big one I know is that PDF is not Turing complete, while PS is.
You don't want your printer running infinite loops or memory bombs
or similar. A malicious postscript document doesn't need a buggy
postscript interpreter to do it's dirty work, a malicious pdf needs
to cause the interpreter to misbehave to do any damage.
Basically, no variables, and no loops or other control constructs. So
there is a one-to-one correspondence between objects in the page
definition and objects on the page rendition.
On Fri, 11 Sep 2026 15:24:23 +0300, magardner2010 wrote:
To convert arbitrary postscript into a PDF, you need to rewrite the
postscript to only use the non Turing-complete subset allowed in
pdf, which I could see running into the halting problem and other
fun stuff.
Adobe had a product called ?Distiller? which did this. You ran a
custom wrapper program on a PostScript interpreter, which I think
intercepted all the actual rendering calls and output corresponding
PDF objects instead.
The big one I know is that PDF is not Turing complete, while PS is.
You don't want your printer running infinite loops or memory bombs
or similar. A malicious postscript document doesn't need a buggy
postscript interpreter to do it's dirty work, a malicious pdf needs
to cause the interpreter to misbehave to do any damage.
Basically, no variables, and no loops or other control constructs. So
there is a one-to-one correspondence between objects in the page
definition and objects on the page rendition.
On 2026-09-12, Lawrence D?Oliveiro <ldo@nz.invalid> wrote:
On Fri, 11 Sep 2026 15:24:23 +0300, magardner2010 wrote:
To convert arbitrary postscript into a PDF, you need to rewrite the
postscript to only use the non Turing-complete subset allowed in
pdf, which I could see running into the halting problem and other
fun stuff.
Adobe had a product called ?Distiller? which did this. You ran a
custom wrapper program on a PostScript interpreter, which I think
intercepted all the actual rendering calls and output corresponding
PDF objects instead.
The big one I know is that PDF is not Turing complete, while PS is.
You don't want your printer running infinite loops or memory bombs
or similar. A malicious postscript document doesn't need a buggy
postscript interpreter to do it's dirty work, a malicious pdf needs
to cause the interpreter to misbehave to do any damage.
Basically, no variables, and no loops or other control constructs. So
there is a one-to-one correspondence between objects in the page
definition and objects on the page rendition.
...then some idiot decided to allow javascript in PDF
On 2026-09-12, Ian wrote:
On 2026-09-12, Lawrence D?Oliveiro <ldo@nz.invalid> wrote:
On Fri, 11 Sep 2026 15:24:23 +0300, magardner2010 wrote:
To convert arbitrary postscript into a PDF, you need to rewrite the
postscript to only use the non Turing-complete subset allowed in
pdf, which I could see running into the halting problem and other
fun stuff.
Adobe had a product called ?Distiller? which did this. You ran a
custom wrapper program on a PostScript interpreter, which I think
intercepted all the actual rendering calls and output corresponding
PDF objects instead.
The big one I know is that PDF is not Turing complete, while PS is.
You don't want your printer running infinite loops or memory bombs
or similar. A malicious postscript document doesn't need a buggy
postscript interpreter to do it's dirty work, a malicious pdf needs
to cause the interpreter to misbehave to do any damage.
Basically, no variables, and no loops or other control constructs. So
there is a one-to-one correspondence between objects in the page
definition and objects on the page rendition.
...then some idiot decided to allow javascript in PDF
This being Adobe, I guess we're lucky they didn't push for e.g. wide use
of Shockwave Flash as part of PDF or something... or that you don't need
a Pantone??? license to see colors...
I can imagine it: "oh sorry your document made the printer run out of
toner because your Pantone White background requires a Pantone license
so it printed as a black page"...
(Cf. <https://merveilles.town/@prahou/117216842053638207> for Adobe Photoshop)
On 2026-09-12 12:07, Nuno Silva wrote:
On 2026-09-12, Ian wrote:
On 2026-09-12, Lawrence D?Oliveiro <ldo@nz.invalid> wrote:
On Fri, 11 Sep 2026 15:24:23 +0300, magardner2010 wrote:
To convert arbitrary postscript into a PDF, you need to rewrite the
postscript to only use the non Turing-complete subset allowed in
pdf, which I could see running into the halting problem and other
fun stuff.
Adobe had a product called ?Distiller? which did this. You ran a
custom wrapper program on a PostScript interpreter, which I think
intercepted all the actual rendering calls and output corresponding
PDF objects instead.
The big one I know is that PDF is not Turing complete, while PS is.
You don't want your printer running infinite loops or memory bombs
or similar. A malicious postscript document doesn't need a buggy
postscript interpreter to do it's dirty work, a malicious pdf needs
to cause the interpreter to misbehave to do any damage.
Basically, no variables, and no loops or other control constructs. So
there is a one-to-one correspondence between objects in the page
definition and objects on the page rendition.
...then some idiot decided to allow javascript in PDF
This being Adobe, I guess we're lucky they didn't push for e.g. wide use
of Shockwave Flash as part of PDF or something... or that you don't need
a Pantone??? license to see colors...
I can imagine it: "oh sorry your document made the printer run out of
toner because your Pantone White background requires a Pantone license
so it printed as a black page"...
(Cf. <https://merveilles.town/@prahou/117216842053638207> for Adobe
Photoshop)
I had a printing system failure the other day, and after some minutes
a page printed fully in black. I think the root of the problem was
that /etc/hosts file was empty.
...then some idiot decided to allow javascript in PDF
On Sat, 12 Sep 2026 07:32:15 -0000 (UTC), Ian wrote:
...then some idiot decided to allow javascript in PDF
That?s not in the ISO32000 spec, as far as I know.
On 2026-09-13 10:16, Lawrence D?Oliveiro wrote:
On Sat, 12 Sep 2026 07:32:15 -0000 (UTC), Ian wrote:
...then some idiot decided to allow javascript in PDF
That?s not in the ISO32000 spec, as far as I know.
It is an adobe extension an only them support it.
Mozilla is trying, but results are not good.
Governments use those documents in interactive forms.
"Carlos E.R." <robin_listas@es.invalid> posted:
On 2026-09-13 10:16, Lawrence D?Oliveiro wrote:
On Sat, 12 Sep 2026 07:32:15 -0000 (UTC), Ian wrote:
...then some idiot decided to allow javascript in PDF
That?s not in the ISO32000 spec, as far as I know.
It is an adobe extension an only them support it.
Mozilla is trying, but results are not good.
Governments use those documents in interactive forms.
With good free alternatives, I've always thought that
government mandating pay-to-play formats like MS Word and PDF
is inappropriate. Am I wrong?
"Carlos E.R." <robin_listas@es.invalid> posted:
On 2026-09-13 10:16, Lawrence D?Oliveiro wrote:
On Sat, 12 Sep 2026 07:32:15 -0000 (UTC), Ian wrote:
...then some idiot decided to allow javascript in PDF
That?s not in the ISO32000 spec, as far as I know.
With good free alternatives, I've always thought that
government mandating pay-to-play formats like MS Word and PDF
is inappropriate. Am I wrong?
"Carlos E.R." <robin_listas@es.invalid> posted:
On 2026-09-13 10:16, Lawrence D?Oliveiro wrote:
On Sat, 12 Sep 2026 07:32:15 -0000 (UTC), Ian wrote:
...then some idiot decided to allow javascript in PDF
That?s not in the ISO32000 spec, as far as I know.
It is an adobe extension an only them support it.
Mozilla is trying, but results are not good.
Governments use those documents in interactive forms.
With good free alternatives, I've always thought that
government mandating pay-to-play formats like MS Word and PDF
is inappropriate. Am I wrong?
Basically, no variables, and no loops or other control constructs.
So
there is a one-to-one correspondence between objects in the page
definition and objects on the page rendition.
"Carlos E.R." <robin_listas@es.invalid> posted:
On 2026-09-13 10:16, Lawrence D?Oliveiro wrote:
On Sat, 12 Sep 2026 07:32:15 -0000 (UTC), Ian wrote:
...then some idiot decided to allow javascript in PDF
That?s not in the ISO32000 spec, as far as I know.
It is an adobe extension an only them support it.
Mozilla is trying, but results are not good.
Governments use those documents in interactive forms.
With good free alternatives, I've always thought that
government mandating pay-to-play formats like MS Word and PDF
is inappropriate. Am I wrong?
James Dow Allen <user4353@newsgrouper.org.invalid> wrote:
"Carlos E.R." <robin_listas@es.invalid> posted:
On 2026-09-13 10:16, Lawrence D?Oliveiro wrote:
On Sat, 12 Sep 2026 07:32:15 -0000 (UTC), Ian wrote:
...then some idiot decided to allow javascript in PDF
That?s not in the ISO32000 spec, as far as I know.
It is an adobe extension an only them support it.
Mozilla is trying, but results are not good.
Governments use those documents in interactive forms.
With good free alternatives, I've always thought that
government mandating pay-to-play formats like MS Word and PDF
is inappropriate. Am I wrong?
I did not look at recent specs, but PDF spec was freely available
and IIUC put no restricions on use. I was able to implement
my little PDF processing utility working from the spec. And
there were several free implementations of viewers and
associated tools. I certainly can do "full" PDF flow, that
is create them, view, convert to bitmaps, add some annotations
using only free tools. That is quite different from MS Word.
The PDF forms extention may be an exception.
The Office Open XML standard(s) were a response to the push that tried
to promote ODF and other open formats on grounds that they were
standards (or at least that seemed to be invoked too).
On 9/15/26 02:41, Nuno Silva wrote:
The Office Open XML standard(s) were a response to the push that tried
to promote ODF and other open formats on grounds that they were
standards (or at least that seemed to be invoked too).
Just say no. I only use OO. Of course I don't have a corporate IT
department bribed by Microsoft trying to declare their stuff as a
standard. People (mostly in Europe ATM) are starting to catch on.
Just say no. I only use OO.
On Tue, 15 Sep 2026 07:41:17 -0700, Peter Flass wrote:
Just say no. I only use OO.OpenOffice??!!?? Why??!!??
<https://www.libreoffice.org/libreoffice-vs-openoffice/>
Lawrence D?Oliveiro wrote:
On Tue, 15 Sep 2026 07:41:17 -0700, Peter Flass wrote:
Just say no. I only use OO.OpenOffice??!!?? Why??!!??
<https://www.libreoffice.org/libreoffice-vs-openoffice/>
Supports more operating systems?
On Tue, 15 Sep 2026 07:41:17 -0700, Peter Flass wrote:
Just say no. I only use OO.
OpenOffice??!!?? Why??!!??
<https://www.libreoffice.org/libreoffice-vs-openoffice/>
On 2026-09-16 01:25, Lawrence D?Oliveiro wrote:
On Tue, 15 Sep 2026 07:41:17 -0700, Peter Flass wrote:
Just say no. I only use OO.
OpenOffice??!!?? Why??!!??
<https://www.libreoffice.org/libreoffice-vs-openoffice/>
Given that table I don't see a reason for the existence of OO. Waste
of good developers time. Unless they have to say some good reasons I
don't know about.
On Wed, 16 Sep 2026 09:19:11 +0200, Carlos E. R. wrote:
On 2026-09-16 01:25, Lawrence D?Oliveiro wrote:
On Tue, 15 Sep 2026 07:41:17 -0700, Peter Flass wrote:
Just say no. I only use OO.
OpenOffice??!!?? Why??!!??
<https://www.libreoffice.org/libreoffice-vs-openoffice/>
Given that table I don't see a reason for the existence of OO. Waste
of good developers time. Unless they have to say some good reasons I
don't know about.
There is one surprising (and disappointing) reason: familiarity.
For some reason, lots of Windows users still remember OpenOffice from
the early days, while the word about LibreOffice has yet to reach
them, even after all these years. So they continue going to the
OpenOffice site in significant numbers, and continue to be
disappointed by the poor quality of what?s available there.
The suggestion has been made that the OpenOffice project should just
shut down already, and turn their domain into nothing more than a
redirect to the LibreOffice site. That would solve a lot of problems.
On Tue, 15 Sep 2026 20:07:29 -0700, Dave Yeo wrote:
Lawrence D?Oliveiro wrote:
On Tue, 15 Sep 2026 07:41:17 -0700, Peter Flass wrote:
Just say no. I only use OO.OpenOffice??!!?? Why??!!??
<https://www.libreoffice.org/libreoffice-vs-openoffice/>
Supports more operating systems?
Nonsense.
| Sysop: | Jacob Catayoc |
|---|---|
| Location: | Pasay City, Metro Manila, Philippines |
| Users: | 4 |
| Nodes: | 4 (0 / 4) |
| Uptime: | 497099:23:27 |
| Calls: | 182 |
| Files: | 744 |
| D/L today: |
6 files (5,132K bytes) |
| Messages: | 73,551 |