X Window System Administrator's Guide: For X Version 11 (1992)
archive.org
archive.org
O'Reilly became famous for the "animal books", which were often indispensable bibles for a software technology.
Today, many of us will attack any problem via StackOverflow or Googling (or laundering open source with LLMs), framed as how to do a very particular something in a stack of software we understand maybe <1% about. (And sometimes we get certifications by studying for a test, as a way to promote our careers.)
But if we wanted to actually understand some level better, such that we don't have to ask as much "How do I do this specific thing, just tell me what to type?", but instead "How does this thing work conceptually, and what are the interfaces?", the documentation we need is usually poor and hard to discover, or nonexistent.
Example: At least one major cloud provider apparently has huge trouble keeping their documentation complete and accurate, and also generates a lot of noise of low-quality documents. (Maybe someone ran the numbers, and found that good docs isn't needed to attract and retain customers at this point of their market position, and poor docs means they sell more support contracts?)
Fortunately, there are still great technical books being published. (Example: if you're an experienced systems programmer wanting to use Rust, IMHO the first book you should read is the Blandy, et al. "Programming Rust" animal book from O'Reilly, and then keep it on your virtual bookshelf forever, for reference.)
I hope that "StackOverflow methodology" and LLMs don't further hurt the market for books and learning, and that we'll continue to see great books from O'Reilly and other publishers.
It wasn't/isn't just a UI toolkit. The O'Reilly series covered all the parts of X, from the drawing primitives, to higher level Xlib, to managing X servers and clients. The equivalent on Windows would be a book series covering GDI, Win32, and configuring every Registry and Group Policy tweak in the Windows shell.
Yeah and that in detail, not just the one line description you get which is often incomplete and sometimes outright misleading.
So, where I worked, as we had multiple licenses and we really only needed to keep one or sets around, teenage me got to take a set home. With that said, my memory is that a large chunk of them was mostly the man pages printed out.
I have fond memories playing the "man game" or the "reverse man game". In the former, player one calls out the name of a UNIX command, and player two needs to recite the one-line description. In the latter, player one reads out the description and player to needs to guess which command is thus described.
While the UNIX manuals (section 3 (API) and especially section 1 (commands)) are well-used, I admittedly didn't use the X11 set as often: the most-used page of the X manual set is the one where it says how to set the DISPLAY variable.
The linked to book appears to be how to administrate an X server. And I know at least one volume covers a higher level toolkit other than Xlib (Athena widget set maybe?)
I didn't write that Xt itself is using Athena, i wrote that the "Xt Intrinsics volume" (i.e. the book) uses Athena as the widget toolkit. Xt itself does not provide any widgets so you need some widgets to show what it does and the book uses Athena for that.
O'Reilly's X Window volumes are not about a toolkit, they document everything that could be relevant to someone about X Windows at the time, including the X protocol itself, the low level X library, a user's guide, the linked administrator's guide and two widget toolkits - Motif and XView - with both references and guides to them.
> once things like VB6 got off the ground.
VB6 also had a ton of documentation, to the point where it wasn't feasible to print it all. VB5 had a 1064 page reference book and a 898 page guide book, among others. Hell, even VB1 for DOS had a 510 page reference book and a 658 page guide book. And these things only discussed a fraction of what the X Windows books did.
We had the whole set in the college library and it took up a whole shelf. I often looked in awe at it, but it didn't seem to get much use.
Of course the waterfall model and lack of online updates made this more feasible. When updates were shipped once a year or so they often came with paper addendums you could clip into the binders.
These days software changes too fast for such documentation to keep up, but I feel we've gone too far, it's often hard to find any clear word on what something does now.
If you build and install all of core+contrib you might need as much as 900 megabytes of disk space. Outrageous!
So we stuck a build of X11 on our servers and mounted it over NFS. It started up a bit slower, but you generally only started an X session once a day (or for some people, once a month), so it was OK.
0: X Protocol Reference Manual
1: Xlib Programming Manual
2: Xlib Reference Manual
3: X Window System User's Guide
3M: X Window System User's Guide OSF/Motif Edition
3O: OPEN LOOK User's Guide (never published)
4: XToolkit Programming Manual
4M: X Toolkit Programming Manual OSF/Motif Edition
5: XToolkit Reference Manual
6A: Motif Programming Manual
6B: Motif Reference Manual
7A: XView Programming Manual
7B: XView Reference Manual
8: X Window System Administrator's Guide
Some of them had multiple editions ("new, for X11 Release 4"). There was also a separate supplement book for... R6? And a lay-flat book "The X Window System in a Nutshell". Oh, and issues of the X Journal, which slowly generated the letter X on their collective spines. All these covered almost a complete shelf.Sometimes I'm amazed by the amount of money we were spending to be able to do our work. On the other hand, I definitely used volumes 0, 1, 2, 4, 5, and 8 a lot!
> Copyright-comment
> Scanned under the Creative Commons Founders' Copyright license, per Tim O'Reilly.
Yes, O'Reilly did, in fact, re-license the content for sharing: