Retarded Ideas in Computing Science

20130904     TerraHertz     http://everist.org

A random list of Retarded Ideas and practices that have arisen during the evolution of computing science.

Each one of these crippled some aspect of Computing's potentials for enhancement of human capabilities and social advance. Each wasted, and continue to waste, many lifetimes of manhours in effort expended solely to work around the stupidity.

See also: </p> is Retarded.

The ASCII 'end of line' coding fiasco.
Creators of ASCII:   "This is for line printers. So we only need CR and LF in the way of line formatting.
Carriage Return to home the print head to the left, and Line Feed to advance the paper."
Microsoft: "Well then, I guess end-of-line in text files will have to be two bytes; CR-LF."
Unix: "What? TWO whole bytes?! Never! We're going to use LF alone for end-of-line."
Apple: "Screw you guys, we're going to use CR alone for end-of-line."
Or to quote http://peterbenjamin.com/seminars/crossplatform/texteol.html
     "They are different in order to protect copyright privileges."
History: "Take them all outside and shoot them. And copyright too."
Tim Berners Lee's idea that HTML would be an entirely display independent scheme, and so should have no ability whatsoever to describe absolute layout positioning in documents. Because if you allowed absolute formatting some display devices might not be able to support it.
Jesus Christ, what a fricking nitwit. And so ever since, html and all its add-on hacks (looking at you, DOM, CSS, Java, Javascript, etc) have been a battleground between the equally nit-wit academics who stubbornly try to maintain TBL's 'purity of abstraction', and the practical people who just fricking want to put this ornate letter T right here on the page. And incidentally the page is A4 sized; 210 by 297 mm exactly. With a torn corner and coffee stain just so, and a few printer's defects in particular letters, though the encoded text must still be fully searchable.

Ultimately that one stupid, stupid misconception doomed html to permanent unsuitability for capturing the content and physical form and appearance of the cultural heritage of humankind. All written works, art, etc in digital form, freely available to all. Which there is an extremely urgent need to do (acid paper disintegration, etc), and which should be one of the highest priority goals of humankind.

To this day HTML still lacks a workable form of the most simple, obvious, indispensible physical format - the page break!
Let alone uniformly sized pages or physical page sizes at all. Also page numbering, footnotes, tab stops, etc.
Tab stops... sigh. I mean what in hell is the problem with being able to define a horizontal position on the page, then easily make text in subsequent lines align vertically with it? And by 'easily' I refer to actually using the ASCII TAB character to do it. Hey look, there's even a key on every keyboard just for this. Tab stops are such a simple, obvious and indispensible feature for document layout. So desirable they've been available on mechanical typewriters since 1897!

But do we have tabs in HTML? Nooooooo... the html 'absolute layout is evil' ideologues insist simple tab stops would be too un-abstracted, too direct. So instead we must all fool around with stupid tedious tables or ridiculously complicated CSS grid schemes.

What a load of bullcrap.
Simple absolute positioning doesn't have to be mandatory you dicks, but the capability absolutely definitely does have to be there.

PS. That's why the content list of this NobLog is the way it is. Each line is a single hyperlink. It would be nice to make the title text align to an absolute position spaced well after the date numbers, without the nuisance of a table or grid. So how?

Adobe: "I know, lets take our beautiful Postscript page layout language, and add 'document security' to it!"
And so we got the abortion known as PDF, and the potentially Internet-killing Trojan software known as Acrobat.
Intel: Segmented architecture. The┃greatest┃waste┃of┃programmer┃time┃ever. Let us never speak of it again.
Microsoft: "Let's put all categories of configuration and state information for everything in the Windows OS, GUI and user applications, all mixed together in one giant central database. But with most of it undocumented. Oh, and we'll break it into arbitrary parts, give them ridiculous names, keep the only current version in memory, make sure there's no easy way to save and restore it as a simple, single compact file, and provide a really crap utility to search/edit/debug it. We'll make the data structures it contains arcane and poorly documented, all the better to hide things (like our deliberate 'look and feel' UI cripples - icon position data, cough.) By obfuscation we'll ensure it always grows ever larger and cruftier over time. Essential linkages will be inscrutible numeric CLSIDs, rather than containing any human-comprehensible descriptive text. The mixing of OS and user-app config data will make your working environment of installed software tools totally uncloneable, thus forcing you to frequently waste hours and days rebuilding it all. The operating system and most applications will critically depend on it, which will make it a central point of failure able to take down a whole machine via the slightest corruption. If that should by unlikely chance ever happen, bwahahaha! Not that we insist you regularly buy a new PC and OS or anything!
We'll call it The Windows Registry."

"We originally wanted to call it the Microsoft BOGU (Benign Object for Greater Utility) but Mr Gates vetoed that for some reason."

In GUI's, making the edges of the screen 'special' in any way. Popup task bar, pull tabs, the Win8 side menu, etc. They are all collossal pains in the butt, due to frequent accidental activation with general cursor movements. There's a MUCH better way (which I'm not going to mention here.)
DRM. This one is so evil, so deliberately evil, so much a tool of the Elites for controlling the peasants and justifying a whole swathe of additional freedom-killing social-control impositions including Internet spying and restrictions, that the absolute best way to deal with this issue would be to take a few hundred of the global financial/corporate Elite outside and shoot them as an example. And keep doing that till the remaining Elite arseholes apologised extremely sincerely and promised to never even think of trying this shit on again.

Here's a short story with an aside about DRM: Fermi's Urbex Paradox.

End User Licence Agreements
"I accept the terms in this licence agreement."
Right. Which I haven't read, don't have time to read, am not going to read under any circumstances, wouldn't understand if I did, and couldn't give a fettid dingo's kidney what the fuck it contains. If you think just because I have to click your little box to make this bullshit go away, that I in any way agree to anything, you're fucking deluded. Listen you shitheads, DATA CANNOT BE OWNED. Period. This is a fundamental consequence of the way this Universe works, probably will turn out to be a law of physics once we dig deep enough, and nothing you say or do can alter that in any way. It's clear that if there really was any original design process for this Universe, the rule 'cannot own data' was likely one of the core customer requirements.

That you corporate arseholes temporarily have a stranglehold on the legal system, and have twisted it to match your insane, perverted ideas about how the universe ought to work, does not make it true, fair, moral, workable or even vaguely sensible.

You keep trying to force this DRM/copyright crap on everyone, which is high on the long list of reasons you're someday all going to be lined up against a wall and shot for crimes against humankind.

Audio CDs - contain no meta data. No track names, artist, album name, album cover artworks, etc. FFS!
XML.
As if html wasn't broken enough, a bunch of me-too wankers thought they'd 'extend and generalize' html. But without fixing any of the fundamental flaws in html of course. The result is an enormous turgid mess, and a lot of exceedingly thick and expensive text books that rarely seem to mention anywhere in the first 20 pages what they are actually FOR. Because if you're too academically abstracted to write a plain English critique of what you are trying to fix by 'extending' something, you're never going to be able to improve anything. All you'll do is make the broken thing more complicated. Which you may think justifies your existence - but on the contrary, you're an idiot, wasting perfectly good oxygen, paper and bits.
Octal numbering in computer systems. Three bit groups, encoding 0 to 7.
Oh God, was it soooo hard back then to see this was a retarded idea? You had a choice: 3 bit or 4 bit groups. Three is a really awkward number to break larger binary words into. But gee, if you chose 4 bit groups, such horror - you'd have to think of symbols to represent the numeric codes 10 to 15. Uhhh.... nope, you couldn't think of anything eh? Complete blank? Too hard?

Dicks. For one thing you missed out on being able to write numbers like FEED DAAD DEAD BEEF.

Well, I suppose there's an upside to the Octal stupidity. I like collecting antique computer hardware. But my loathing of Octal is enough to keep that rather expensive, time and space consuming and ultimately pointless hobby in check.

New computers with no system setup CD.
It used to be that all new computer systems came with separate physical media containing everything needed to reinstall the Operating System from scratch. Which is essential, if your hard disk dies and has to be replaced.

New computers typically no longer come with install CDs. Instead they include 'recovery partitions' on the hard disk.
This is a hidden partition, that the system's BIOS can be told to use to restore the OS.
But... hard disks do fail. More to the point the hard disk command set also allows the drive firmware to be rewritten - which means that it's possible for software on the system to brick hard drives, by overwriting the drive firmware with garbage.

So let me get this straight. The situation now is that virtually no one has a physical backup installable copy of their computer's operating system and drivers stashed away in a cupboard. Meanwhile, almost all PCs are internet connected virtually all the time. The NSA has 'weaponized' the internet backbone, in ways that allow them to gain backdoor access to pretty much any PC they want. Also, everyone runs software like Adobe Acrobat and Flash, which frequently phone home to fetch 'security permissions' and also have potentially full access to everything in the computer. (That's PCs, Macs, and most everythingg else, since there's either Acrobat and/or Flash on just about everything.)

So lets just imagine for a moment that some crazy elite-run government wanted to 'Off the Internet' AND most computers in the hands of the civilian population all in one go. "Why would they want to do that?" you ask? Seriously? Have you been paying attention to politics at all? Anyway, supposing that was what they wanted. As things are now, commands could be broadcast to (say) the Flash player in every online PC to brick the system's hard drive.

And then... even if you had a blank replacement hard drive, you couldn't get your system up again because you don't have a system install disk (CD, or whatever.) Even if you have a copy of Windows (unlikely these days with Win 7 & 8) you don't have the drivers for your particular hardware. And the Net's not working... And your neighbour's PCs are all bricked too...

When people talk about a government controlled 'Internet off switch', I don't think they are really thinking through how thoroughly we'd be screwed, if someone actually decided to do it. Because the almost total absense of physically separate reinstall media these days has put us in a position that is very much like a guy hanging by one rope, several hundred meters up a cliff. If the rope gets cut there's no going back.

I was reminded of this situation recently when I needed to make a bootable complete copy of my main PC's system hard drive, before attempting to alter the configuration of the RAID array that the system boots from.
Oh that's right, I forgot. WinXP has NO facility for creating a bootable primary partition on a drive. It's only the Windows install CD that has that ability, since FDISK and SYS were dropped. After they failed to be updated to match larger hard drive sizes.
But later Windows versions installed on PCs don't even come with CDs.

Document compression schemes using patch tokenization.
This is one I'd never even imagined could happen, till I came across these articles recently:
Ref: Number mangling with image compression
JBIG2 image compression scheme substitutes patches, leading to actual numerical errors in documents.
2013-08-02 http://www.dkriesel.com/en/blog/2013/0802_xerox-workcentres_are_switching_written_numbers_when_scanning
2013-08-08 http://www.dkriesel.com/en/blog/2013/0808_number_mangling_not_a_xerox-only_issue

Ah. Now he explained it I understand why a lot of scanned old technical manuals I see have this problem severely. With actual corrupted text, caused by exactly that problem - incorrect patch substitution.

The whole idea of scanned text compression by tokenizing and fuzzy-matching fairly large areas of the image, then substituting 'close enough' equivalent patches in the text is one of the more retarded concepts I've seen in recent years of computing science. I can't believe anyone ever thought this could be a good idea, or didn't notice the screwed up document scans that result.

Here's an actual extreme example of this stupidity.