Since we're all admitting to have done this at one point in the past, here's mine; PSTab, a guitar tablature typesetter I wrote in 1994. The code is...not great.
http://www.ctan.org/tex-archive/support/pstab
I heard from a guy in NZ a while after that he'd used it to typeset a book of banjo music - it was the only software he could find that supported anything other than 6 strings.
PDF displaced Postscript for two reasons: (i) lower resource usage, and (ii) better handling of fonts.
Since then PDF has become more sophisticated, handling metadata, tagging, and embedding media. It's been a long time since Postscript has been "the universal language for graphical and printed work".
Not exactly. A PDF is generated from PostScript, as noted in the article:
"A PDF file is a static representation of the result of executing the PostScript file on a single occasion."
A PDF is a more compact version of a PostScript file - we could call it compiled PostScript, to use a term many on this site are familiar with. Just as a compiler has many output options, distilling PostScript can result in many different file sizes and levels of quality. (Check out the Distiller output settings.)
No, PDFs don't have to be generated from PS at all. PDF has it's own well-defined spec. Valid PDF files can be generated in many ways besides from PS files. It's even possible to write a valid PDF file by hand if you're feeling masochistic.
PDF is typically directly generated these days. There's no need to generate it from PostScript first.
The main difference between the two is that PDF is a declarative definition of the page whereas PostScript is a turing complete language. You can write a PostScript ray-tracer which will execute on the printer itself something which would never be possible in PDF. On the flip side, PDF is far more efficient as an electronic document format.
My first job out of college, I had a Turbo C compiler, a DEC PC, a database app (government) written in Foxbase, and a Postscript printer. No real reporting tools, so all my reports and add-ons were written in C and outputted Postscript directly to the printer. Was a very interesting learning experience.
A few years ago I took a look at this Unix circuit schematic editor, which seems to be still going strong. Symbols are kept directly in Postscript, if memory serves.
http://opencircuitdesign.com/xcircuit/ : "XCircuit regards circuits as inherently hierarchical, and writes both hierarchical PostScript output and hierarchical SPICE netlists."
It'd be really cool to have an ebook reader that converts epub files, and runs either TEX directly or postscript implementing the TEX paragraph formatter.
It's really bugging me that my kindle is incapable of running Knuth's algorithm, preferring instead to avoid hyphenation entirely, serving up an ugly hodgepodge of justified and ragged-right text. C'mon guys, this is old stuff...
When I wanted to investigate fractals using Scheme, I had the Scheme program spit out PostScript output, then ran it through gs to look at the results.
PostScript is really tractable as a language to work directly in. Basic line-drawing stuff is almost as simple as the Logo I played with as a kid.
I'm not sure I'd ever try it myself — for one thing, PostScript would only be useful for preparing documents destined for print, not HTML — but if I ever did need such a thing, I'd be interested to try it out.
Whether the rasterization in the end targets a printer or it ends up in a .png doesn't really matter at all.
But yeah, sadly there's no readily available system nowadays with true Display PostScript. Not very likely that Oracle will release the NeWS sources, and OS X switched to DirectPDF.
I like hand-coding PostScript. Been doing it since the early 90's. It doesn't need to be a mess and can be beautifully concise and clean. It does have some odd ideas for the beginner (the page always starts at the bottom left coordinate and goes up, not down) but that's not hard.
I remember back in the days of "early" Linux, when printer drivers were still a bit rare, there was lots of software that generated somewhat readable PS, i.e. the programmers wrote it and then used it as a template, like you'd be using HTML templates nowadays. The timetables for a German railway info site did that.
And I made heavy use of Jamie Zawinski's audio-tape.ps [1] back then. Never got through Thinking in PostScript [2] completely, though…
Wow. I'd completely forgotten about audio-tape.ps. It had all sorts of nifty built-in graphics for various bands, and I remember learning way more than I ever expected to want to know about PostScript adding similar customisations for others.
I like to write Perl scripts which generate Postscript pictures. It gives me better interactivity of Perl development combined with the exactness of the results of Postscript.
The end result is usually a PDF file.
Still I've never learned to embed external fonts in my scripts, does anybody have some straightforward text about that part? As far as I know I'm legally 100% clean if I manage to use any font as long as I only distribute the PDF at the end.
It's been a long time since, but as far as I remember it basically boils down to pasting the pfa into your file (pfb2pfa is your friend here).
And it totally is a murky legal situation, many fonts don't allow embedding into a pdf, and you may not use almost all fonts in any way at all if you do not have a valid license. Fonts in the abstract may or may not be copyrightable depending on where you live, but the actual pfa-code you put in your file and the representation generated from that when converting to PDF definitely are.
You might want to try cairo for that instead. It has PS and PDF backends which are pretty direct (see the bottom of this page http://cairographics.org/documentation/using_the_postscript_...), and it'll take care of font subsetting and embedding for you, if that font allows it.
You need the T1ASCII utility to convert your binary Type1 PS Font to a Type3 ASCII font. You can then embed the font as a PS resource and call it from within your PS file as you would normally.
T1Utils is the package and can be found from your repository or synaptic. There's also a DOS version and it comes with its c code.
Nice. FWIW, We emit "direct" postscript for most of our generated warehouse documents like packing slips, gift cards, snail-mail gift certificates, invoices, etc.
Pie menu widget: http://www.donhopkins.com/home/code/piemenu.ps.txt
PizzaTool graphical Pizza ordering gui: http://www.donhopkins.com/home/code/pizzatool.ps.txt Illustration: http://www.donhopkins.com/home/catalog/images/pizzatool.gif
VT100 terminal emulator: http://www.donhopkins.com/home/code/newterm.ps.txt
UniPress (Gosling) Emacs display driver: http://www.donhopkins.com/home/code/emacs.ps.txt
Metacircular PostScript interpreter: http://www.donhopkins.com/home/code/ps.ps.txt
PSIBER Space Deck (visual PostScript debugger and programming environment): http://www.donhopkins.com/home/code/litecyber.ps.txt More about that: http://www.donhopkins.com/drupal/node/97
A lunar lander game James Gosling wrote in 1988: http://www.donhopkins.com/home/code/lander.ps.txt
HeapSort by Owen Densmore: http://www.donhopkins.com/home/code/heapsort.ps.txt
QuickSort by Don Woods (an incredibly twisty maze of stack manipulation, by the Adventure dude): http://www.donhopkins.com/home/code/quicksort.ps.txt
BubbleSort by Bobo Leffler (he was afraid this might come back to haunt him): http://www.donhopkins.com/home/code/bubblesort.ps.txt