My first "desktop publishing" experience was using troff on Unix and sending output to the university's phototypsetter. Then, in the second half of the 1980s, I worked at a Fortune 200 company in the team rolling out PCs and networks across the company's many locations.
At the time, the basic choice was the HP laserjet with optional font cartridges (which were very expensive as I recall, something like $200 each in 1980s dollars), or the Apple laserwriter which was something like $4K if I recall correctly, but had Postscript built in, so no need for hardware font cartridges.
So as a result, the HP hardware got rolled out to most offices where mostly what was needed was to print memos in default monospace font, while to make a nice looking newsletter, flyer, or brochure, you had to travel to headquarters to use the Apple printer.
My first printer was an HP LJ II+ with the PostScript cartridge. I had to buy extra RAM for it and for a while, I had more RAM in my printer than in my computer. A few years later, I bought a fancier PS printer that printed ledger size pages and handled duplex and again I had more RAM in my printer than in my computer. I used to occasionally write PS documents by hand (one was to print out country placards for the Model United Nations club at my college where for country names which were especially long, the text would be auto condensed.
These days, with the advent of printing subsystems in the GUI and the fact that software no longer needs to handle the printer directly, I don’t know if anyone ever buys PS printers at all anymore.
PDF, although it can resemble PostScript, is not PostScript. PostScript is a Forth-derived Turing-complete programming language that, in an era of generic printer drivers, offers little benefit anymore since it’s unlikely that an OS printer interface would take advantage of its capabilities. In the old days of programs managing printer output directly, it was handy to be able to do things like decompress TeX PK font bitmaps on the printer (or even more dramatically to send binary data for illustrations to be decoded and decompressed on the printer), but the set of printer primitives that are supported by OS printer APIs is constrained enough that nobody will do these sorts of things anymore. PDF is now essentially the lingua Franca of print interfaces and the programmability of PostScript is more a liability than a benefit.
I was delighted to discover a full chapter dedicated to troff in a textbook I picked up on Berkeley Unix. Raw documents written in troff (the ones that I’ve seen) seem so quaint. Like little homemade chicken pot pies.
I’m mixed on the state of desktop publishing today. I’m missing that quaintness.
I wasn't aware there was a state of desktop publishing today. DTP has mostly been superseded by the web. DIY amateur print of posters and such is more likely to be hammered together in Word than in Indesign.
Design agencies still do design-for-print, but there's usually an integrated workflow of some kind which means web and print assets - including styles - can be shared and reused.
It's a long way from the blocky output of a 300dpi laser in the early 90s assembled on a PC running a $50 DTP suite.
The sad thing about Postscript is that it could have been the foundation of the web, which would have made web and print integration much simpler and design more straightforward and consistent.
Instead we got the sprawling incoherent catastrophe that is HTML/CSS/js for the web, PDF for most other document-ish things, and Epub (which is a dumbed down and DRM'd version of HTML etc) for digital book publishing.
Early in my career, I worked at a newspaper. Pages were laid out by literally cutting pieces of paper and gluing them (with semi-adhesive wax) onto big sheets of paper. Sometimes on a backlit table, so you could see what the camera would see. Then it was shot with a huge camera. The remainder of the process was messy and entirely mechanical.
“Desktop publishing” changed ALL of that.
And now, we just call it “print.”
Postscript’s contribution to this transformation can hardly be overstated. To have a language for describing a printed page that could be exchanged between all the disparate systems of the 90s, …it was a revolution.
> Pages were laid out by literally cutting pieces of paper and gluing them
Cow Gum, it was called in the UK (I don't know about elsewhere). I was a graphic designer in the late 1980s, and remember all that analog equipment, including Rotring pens (insanely expensive) and line tape, and photomechanical transfers (PMTs).
Back in early 90s I was working for a big international company using VAX servers and DEC laser printers in their engineering department. The wanted to have various types of graphical reports outputted based on data in their INGRES database. To produce those graphical reports I used FORTRAN and hand-written PostScript code. Did a fair amount of PostScript coding back then.
Remember I purchased a set of PostScript books to study. I still have them in my library:
I was disappointed when Postscript was initially released, as they were at pains to port the constraints of paper to the computer, which I considered (and still consider) a gross step backwards. Display postscript (Sun and NeXT mostly) was an improvement but documents were still stuck in a paperish mode. Few PDF are truly editable.
This was made much worse when displays migrated from portait by default to landscape by default.
Even today my heart sinks when I see that I've been given a pdf
The portrait format of most PDFs is one of the reasons I always insist on having a portrait mode monitor.
My current setup uses three 4K displays: my ThinkPad P1's built-in display plus two 24" 4K monitors (currently LG, but I've also used Dell).
One of the 24" monitors is above the ThinkPad display, in landscape mode with the left edges aligned. (Actually the monitor's left edge is just a pinch to the left of the ThinkPad display's, so they look aligned when I'm using them.)
The other 24" monitor is to the left of these, in portrait mode. The bottom edge is about 3/4" above the desk, allowing room for cables to go underneath it.
The portrait mode display is perfect for reading PDF files, and also great for HTML documentation, long text files, etc.
Note that 27" monitors would be less ideal for this configuration. The top of the portrait mode 27" display is too high up and would cause neck strain.
All three displays are the same distance from my eyes, about 20", and I wear single-vision prescription glasses adjusted for that distance. The ThinkPad display scaling is set to 300%, and the 24" displays at 200%. It is a pretty great setup.
It is a shame. My first encounter with PostScript was getting a TeX / LaTeX post processor to output PostScript to the universities newly purchased LaserWriter (so much CamelCase ha ha).
When the Sun rep brought us a tape with NeWS to load up to our Sun Workstations we were quite excited, but it was so slow it was nearly unuseable. We scurried back to SunView pretty quickly.
edit: can't remember if it was dvi2ps or dvips, but we ended up having to hack it to be able to embed graphics because LaTeX at the time either couldn't do that or could do it but dvi2ps didn't understand it. Fun times.
>This was made much worse when displays migrated from portait by default to landscape by default.
When were displays ever portrait by default? I'm only familiar with resolutions wider than tall. The oldest I've used were the amber/green monochrome CRTs. I honestly don't know the dimensions of those but they were back when 80chars per line (or whatever the count) limits existed. But even then, it always seemed to me to be more chars across than rows down.
Character terminals like the AAA were portrait — 80x48, yes, but characters were taller than wide.
But I was thinking if bitmapped displays since we were talking about PDF: the Alto was the first such, and was bitmapped, as were CADRs, BLIT, PERQ etc…
The early workstations like the Alto, PERQ, and CADR were portrait by default, and used specialized CRTs.
Personal computers tended to use 4:3 displays because they started out with either terminals (which used television CRTs) or by being plugged into literal televisions.
Workstations went to two-page displays a decade and change before even 1024×768 was widespread for personal computers outside ones used as workstations (e.g. very high end PCs and Macs used for CAD, page layout, etc.).
Postscript was solving the problem of producing nice proofs for photo transfer. Complaining about how it's not a great display protocol on modern flatscreens isn't missing the point, it's not even realizing there is a point to be missed.
I used the Dover and its postscript predecessor (I can’t remember the name) so I’m familiar with the page description languages described in the article (I was at PARC for a while). PDF was supposed to go beyond that, and its parametric model was a conceptual win…then mostly constrained as a semantic-free rendering system.
You would say the same thing if you were trying to edit a Word document in a browser and you'd never seen a word processor. All PDF can be edited like any other electronic document, you just need the right tools. There are a bazillion PDF editors now and Google sucks worse every day, but I learned to edit PDF with PitStop Pro. I could edit a PDF as though working with QuarkXPress on a Quark document or working with Adobe InDesign on an InDesign document. Once you have access to internals of a PDF and tools to manipulate the objects and type, editing PDF is no big deal. Few locks can be truly opened if you don't have a key (or a pick, or a hammer, etc.).
Hmm, yes and no. If you want to be able to edit more than a single word or two, I'd rather have the Word doc, it can at least handle wrapping text, widow control etc. (even if the format is a moving target and even MS can't seem to always handle it... so yes, in 20yr you'll need to do your edits via DOSBox). If you only need to read it, PDF wins certainly.
PostScript have followed some previous ideas but the most important part is a thing lost in modern software: a programming language to create documents AND applications at the same time. A kind of document UI years after Xerox.
Modern publishing software have ditched this approach for another, a fully visually one, formally easier, sometimes easier, but in general limited and limiting.
HTML is just a markup language, PS is a programming one, NeWS [1] was an X11-alike server PS-based, SUN Pizza Tool a nice demo [2] oh, sure in term of ease of use is a bit awful to write and read, but still it's a "graphics oriented programming language" while HTML just target crafting clickable text with limited extras.
Kind of sad to see PostScript slowly fade away these days. My Mac won’t even open .ps files anymore! The entire graphics stack was designed around it and they removed viewer support for it…
https://news.ycombinator.com/item?id=33820907