This is triply silly. If the Guru had read just a teensy bit of linked list literature, surely he would have seen Nick Wirths way to link both ways with just one pointer.
It's only been around since, like 1972
This is triply silly. If the Guru had read just a teensy bit of linked list literature, surely he would have seen Nick Wirths way to link both ways with just one pointer.
It's only been around since, like 1972
I've been oracle-free for two years now, so I've forgotten 95% of the headaches. The best ones I can remember:
The Java installer. It sucked really bad. First
off it wouldnt run at all with the standard version 2.37 of Java, you
had to download an install a 11MB patch to Java first. Then it
would run, but all I remember is it asking a lot of unanswerable
questions, informing me my disk with 23Gb free didnt have enough space,
as some integer overflowed and it thought I had -1433GB free.
---
After finally getting it installed, it was time to make a test
database. You'd think by now you could type "dbnew foo
/tables=... /fields=....". Nope.
There's a horribly convoluted 57 pages in the manual as to how to
get a db started, and a lot of handwaving. Crazy.
Particularly annoying was having to generate all kinds of
tables for Oracle's internal use. When you go
to buy a jar of Heinz ketchup, do they require you
to fill out a purchase order for their tomato squisher?
When you buy a car, do you have to go order them some iron ore?
I kept looking and looking, assuming I was mssing the little blue button labeled "make me a database in 2 seconds". Just couldnt seem to find it.
Finally I sat down and wrote some code to do all the necessary
things in one fell swoop. Why wasnt this done on day
one of version 1.0? Yes, I know, there are some
tools that allegedly do this, but they looked far to horrible to ever
use.
That was just the tippy-tip-top of the iceberg. All else has faded from memory. Whew.
if we were still in a cold war with the russkies, we should have dropped Oracle on them.
Free CD's. Please use.
There's a lot of kludges in this code. Some of the more obvious ones:'
(1) The rng function returns a double, but it can only return a bit over two million diferent values, wasting half the precison of a typical double.
(2) then it multiplies the 0..1 value by the wrong scale factor. To get a number between one and x you want to multiply the 0..1 rng by (x-1), not (x+0.99).
(3) Then since the result can be somewhat out of range, there's the kludge of dividing by ten if it's too big.
(4) And also the strange business of scaling up results under 0.01.
Just plain badly thought out code, then badly patched to boot.
Oh you youngsters with your hi-falootin' password obscuration methodologies!
In the OLD days we used Model 33 teletypes to log into the computer. The better systems could turn off the character echoing to ask for the password, but some crude systems could not. So what did they do:
PASSWORD:
MMMMMM\rIIIIII\rOOOOOOO\rIIIIIIII\rAAAAAAA\rWWWWWWW
Yep, make a black smudgy area seven chars wide that you'd type your pw over.
Worked fine on a TTY 33, not so good on a CRT terminal.
The real WTF:
IMPORTANT:
Because fraudulent ("phishing") e-mail often uses misleading links,
Microsoft recommends that you do not click links in e-mail, but instead
copy and paste them into your browsers, as described above.
... but of course copy and paste isn't enough to stop a phish. The instructions would have to get a bit lengthy with phrases like "now inspect the URL for any elements that could be a problem. Perhaps the domain name www.godaddy.microsoft.com isn't really owned by microsoft. Perhaps the letter o in www.microsoft.com is actually the Azerbaijaini sanskrit Unicode for their vowel "owww". Perhaps the URL looks fine up front but in the end redirects to pirates.ru. Yes, learn to be a internet detective and inspect every URL for many minutes before following any link.
It even has a performance bug-- the loop needs only to go up to strlen(a) - strlen(b)
Hmm, there's no need for an array, or sorting it.
Just ping each site in turn, if the price is better than the last best price, remember this site..
No arrays, no sorting.
Perhaps they wanted to bottleneck all the tests against null in order to:
(1) Sometime in the future count all the comparisons against null.
(2) If necessary add some patch in there where the object might be not null, but equivalent, such as the empty string, or "NULL" , or some other data that might have sneaked into the data
(3) As a quick thing to eventually change and get a slight performance boost. :)
What the he ?? I've never seen a font ose one of its characters. Maybe they squeezed the font coser together too much and the narrow characters got kerned out of existence?
eh, it might be one of several things:
(1) the tapes are 100GB "unformatted". Once you lay down inter-block gaps, sync bts, preambles, headers, and CRC's, you could easily lose 10 to 20 percent.
(2) Tapes are known to be a bit unreliable. A good backup program would be smart to write out ECC data to recover from errors. That ECC data takes up extra space of course.
Don't EVER alias things this way!
Think of what happens if....
(1) Your .bashrc file gets deleted.
(2) Your .bashrc file is on a file server that hiccups.
(3) You come over to my computer and do a "rm *.c" expecting it to ask you about each one.
(4) You somehow end up in the Bourne or Korn shells that don't read up .bashrc.
Bad vibes....
This isnt all that far off-- the constants will default to ints, which may be too narrow to tolerate all that left shifting.
one WTF is all the careful left-shifting of 0x0.