Hm. How about, then, the 'blessed binary' method with public-key authentication? For all you non-[ex-]Netrek players, Netrek servers generally bar non-blessed clients with a crypto challenge, so that modified ('borg') clients get bounced.
I believe that it has been worked around with a proxy and some cleverness, but it complicates the matter and does *not* require that the client have a known buffer overflow problem.
Network clients do not have any business accepting more data than they can handle.
* "Encouragement" is not necessarily the same thing as easier admissions.
* Admission into CS does *not* guarantee graduation -- as CS, or anything else.
It's not like the various profs and TAs will suddenly decide that "Oh, we've got to have parity in our distributions, and therefore have mercy in grading under-represented groups" -- be they female, African-American, clones of Mr. Ed, sentient tripedal fungi from Pluto, yadda yadda.
If somebody manages to get admitted, but finds out that he or she is unable or unwilling to walk the walk, it'll likely show.
Encouragement? There's not that much wrong with encouragement, methinks -- that's far different from (mine, as well) the university setting different acceptance criteria.
If they wish to raise their profile, and believe that for some reason female high-school students are less likely to apply despite their qualifications, then fine. More power to 'em. No complaints here...
Now, what *would* be more debatable would be if they either established different criteria (such as [completely arbitrary and random example] increasing verbal but decreasing math score requirements on the SATs), or specifically and systematically gave females an edge in admissions for no reason except their gender.
As long as they do not, and maintain the overall integrity of the system, fine.
Another possible factor: Was the U of Oregon getting any heat from the software or music companies? Rumor says that the music industry has been in occasional contact with people at my own, because the local Windows share network happens to be pretty rife with illegal MP3s, and so forth. If an industry contacts the U's legal people, they might have had more motivation than to just say something like, "We've investigated you because of the unusual network load to your machine. Please cease and desist, or else we may have to enact additional penalties ranging from loss of network connection to expulsion."
Check your license agreements. Some explicitly allowed you to make a single copy for backup/archival purposes; others do not, and implement methods to stop you (such as the old EA schemes involving burning a specific sector, et cetera) while usually offering a set of backup media for $10.00.
* Office applications. Off-hand, are there any GPL/BSD/etc-licensed [including free for *commercial use*] office suites that provide both easy migration, and full functionality?
* Game development -- not so much as engines, but data. There are projects that take a LONG time to develop and write, and potentially a LOT of resources.
At 10Mbps Ethernet, and given access to CD burners, it pretty much *will* happen. It certainly happens here, where students have occasionally attracted the attention of our sysadmins for serving extreme amounts of traffic -- full-length movies (e.g. _SW:TPM_), applications, and so forth.
Try Cygwin, which apparently will serve ('tho it's NOT free if you are porting commercial applications...). According to its FAQ, it now supports both their Unix-like API and the native Windows foo.
It would not surprise me at all if they trademarked 'Wizard' in that context, since I seem to remember competing programs having parts with identical functions but somewhat different names (not an unusual event: remember Borland's DAD vs. MS's MOM? That's as silly as Unix naming traditions...).
So call 'em Assistants, Advisors, Aides... or if you want to tweak MS at the cost of obscurity, Conjurors.
I'm actually a tad surprised they never went after StarDivision, given that StarOffice (at least early versions... 3.0-ish) looked very, *very* similar to MS Office in many, many details.
'ccording to the log, they were never released. Which, frankly, disturbs me based on a quick reading of the log (mind you, I'm not a web admin) -- here's to hoping that the bug fixes in those versions weren't for anything frequently tripped.
If that 1.3 Changelog is correct, then all previous versions of Apache had memory leaks in various places as well as a distinct segfault problem (since these were fixed in post 1.3.6 versions). That would strike me as odd, 'tho: perhaps they released intermittent fixes?
Sounds like something the various distro maintainers should obtain ASAP, at least if they expect customers who wish to run decently loaded HTTP servers with fewer worries.
You're probably mixing up "WindowMaker" -- a very specific window _manager_, ala fvwm, twm, mwm, blackbox -- with "GUI program designer" or something like that.
VB has naught to do with window _managing_, which involves placement of windows, title bars, widget sets, and so forth. WindowMaker has naught to do with writing programs, any more so than any other WM.
To answer your Q 'tho, well: get 'Learning Perl' and it's not that bad. For GUI, consider Java or Tcl/Tk if you must have portability, and there are probably other candidates, too. C w/ XForms, 'tho, is *fast* for programs that require X...
Hmmm. How many mobos would Asus need to sell to make a new design profitable?...and can AMD guarantee a large enough production run?
The one thing about Intel is that it _will_ sell. Apparently, they can toss in processor ID weirdness and a mostly-incremental set of improvements and practically create another brand, and it will _still_ sell like mad. They could probably paint their CPUs orange, top 'em off with purple heat sinks shaped like drunk, lustful goats, and _still_ sell 'em -- in vast quantities, even if leading reviewers publicly question Grove's sanity.
AMD's ability to pump out a large volume has been repeatedly questioned. Not its ability to generate demand -- there's enough of that at the techno-geek level, most likely -- but to fulfill it...
Aye. As long as you've got all the needed libraries with no conflicts (I *think* that dyn-linking an X11R6 application with libc6 and X11R6 libraries dyn-linked with libc5 can do evil things, but it's been a while...), which is increasingly rare, you're fine.
Apps be distributed with the various libraries, or statically linked. This is the same issue w/ Windows, only somewhat better (in that DLLs tend not to have version numbers in their names, IIRC, and thus this makes it hard to have multiple versions...).
I've used Visual C++ on an NT box, but in general it makes me a tad less happy then I am devving on a Solaris or Linux box w/ XEmacs, make, and gdb... but that's probably largely because of the work I normally do.
I loathe the project configuration/maintenance in VC++: give me Makefiles (and not the strange ones that VC++ exports, which in my experience A) usually don't work w/ nmake.exe, and B) are not exactly designed for readability or re-usability).
The other major beef is more with MSDN than the IDE: the documentation seems a lot less... functional than the man pages that I'm used to for Solaris and Linux. I'm used to seeing, for instance, mentions of which header files and libraries to reference for a particular function; any implementation bugs regarding behavior versus existing standards; and so forth -- and I'm used to man-page-style documents being reasonably *up to date*.
That's not to say VC++ has no redeeming features; it does. But if you intend to port to other OSes, or don't have (paper) Win32 programming references around, or so forth -- you might be better off with a different package.
Re:No X -- we need a media-savvy, compositing GUI
on
Is X The Future?
·
· Score: 1
Hmmm. What's so difficult about looking at the technical specs section of a monitor manual? Reading? Searching the table of contents? The actual typing?
Heh. Why use a microkernel? Go ask Torvalds, or read his response to a fairly pro-monolithic-kernel chap in the Letters section of a recent "Communications of the ACM" (don't have the issue w/ me; it's the one that has a special section on lifelike graphics). It's a pretty interesting read.
That's basically the conclusion that I drew: to stop interception, you must be able to use a format that nobody can read unauthorized, but current destinations (audio devices) weren't exactly designed that way.
How do you send a "secret" message to somebody who does not know any encryption algorithms and cannot keep *any* secrets (including those in his head)? This is an analogous problem: the audio signal is *everything*, but also happens to be necessary -- and it's basically cleartext from a pirate's POV.
As somebody else noted, require special hardware on the receiving end, as hardware reverse-engineering would make this a lot more difficult, most likely.
You're not allowed to replace the audio device interface.
The interface only accepts normal, raw audio data.
It is possible, by virtue of the operating system design, to capture data going to the interface.
then it can be intercepted, it would seem. So, break one of the assumptions: require audio devices that can handle the protected data, or block all forms of interception (which might be tricky if one includes attacks on the physical connection?). Ugh.
Well, perhaps there's a more elegant way, but it's not as obvious as a giant mutant glowing crow in a snowfield.
Hmmmmm. I'm not a moderator, but perhaps there should be a "DIE THREAD DIE" option. Or some other restriction on obvious cascades, which seem to be either
a) two camps with immutable and opposing opinions battling each other on an issue, or
b) (today at least) pointless content-free cascades.
Hmmmm. I see mentions of subscriptions on SuSE's (English) site, but their ordering page doesn't seem to list subscriptions as an option -- just their assorted one-time-purchase packages. Odd. Perhaps that's just lagging a bit?
Sure they are (undercut by $2 CDs); you can still burn 'em as long as you read the instructions about what you don't include (i.e. you'll still get a working distro).
Slashdot Mirror
Hm. How about, then, the 'blessed binary' method with public-key authentication? For all you non-[ex-]Netrek players, Netrek servers generally bar non-blessed clients with a crypto challenge, so that modified ('borg') clients get bounced.
I believe that it has been worked around with a proxy and some cleverness, but it complicates the matter and does *not* require that the client have a known buffer overflow problem.
Network clients do not have any business accepting more data than they can handle.
Oh, give the system a little credit, eh?
* "Encouragement" is not necessarily the same thing as easier admissions.
* Admission into CS does *not* guarantee graduation -- as CS, or anything else.
It's not like the various profs and TAs will suddenly decide that "Oh, we've got to have parity in our distributions, and therefore have mercy in grading under-represented groups" -- be they female, African-American, clones of Mr. Ed, sentient tripedal fungi from Pluto, yadda yadda.
If somebody manages to get admitted, but finds out that he or she is unable or unwilling to walk the walk, it'll likely show.
Encouragement? There's not that much wrong with encouragement, methinks -- that's far different from (mine, as well) the university setting different acceptance criteria.
If they wish to raise their profile, and believe that for some reason female high-school students are less likely to apply despite their qualifications, then fine. More power to 'em. No complaints here...
Now, what *would* be more debatable would be if they either established different criteria (such as [completely arbitrary and random example] increasing verbal but decreasing math score requirements on the SATs), or specifically and systematically gave females an edge in admissions for no reason except their gender.
As long as they do not, and maintain the overall integrity of the system, fine.
Another possible factor: Was the U of Oregon getting any heat from the software or music companies? Rumor says that the music industry has been in occasional contact with people at my own, because the local Windows share network happens to be pretty rife with illegal MP3s, and so forth. If an industry contacts the U's legal people, they might have had more motivation than to just say something like, "We've investigated you because of the unusual network load to your machine. Please cease and desist, or else we may have to enact additional penalties ranging from loss of network connection to expulsion."
Check your license agreements. Some explicitly allowed you to make a single copy for backup/archival purposes; others do not, and implement methods to stop you (such as the old EA schemes involving burning a specific sector, et cetera) while usually offering a set of backup media for $10.00.
Hmmmm. Two particular areas.
* Office applications. Off-hand, are there any GPL/BSD/etc-licensed [including free for *commercial use*] office suites that provide both easy migration, and full functionality?
* Game development -- not so much as engines, but data. There are projects that take a LONG time to develop and write, and potentially a LOT of resources.
At 10Mbps Ethernet, and given access to CD burners, it pretty much *will* happen. It certainly happens here, where students have occasionally attracted the attention of our sysadmins for serving extreme amounts of traffic -- full-length movies (e.g. _SW:TPM_), applications, and so forth.
Try Cygwin, which apparently will serve ('tho it's NOT free if you are porting commercial applications...). According to its FAQ, it now supports both their Unix-like API and the native Windows foo.
It would not surprise me at all if they trademarked 'Wizard' in that context, since I seem to remember competing programs having parts with identical functions but somewhat different names (not an unusual event: remember Borland's DAD vs. MS's MOM? That's as silly as Unix naming traditions...).
So call 'em Assistants, Advisors, Aides... or if you want to tweak MS at the cost of obscurity, Conjurors.
I'm actually a tad surprised they never went after StarDivision, given that StarOffice (at least early versions... 3.0-ish) looked very, *very* similar to MS Office in many, many details.
Try Apache's own notes on performance, by Dean Gaudet. Hope that helps.
'ccording to the log, they were never released. Which, frankly, disturbs me based on a quick reading of the log (mind you, I'm not a web admin) -- here's to hoping that the bug fixes in those versions weren't for anything frequently tripped.
If that 1.3 Changelog is correct, then all previous versions of Apache had memory leaks in various places as well as a distinct segfault problem (since these were fixed in post 1.3.6 versions). That would strike me as odd, 'tho: perhaps they released intermittent fixes?
Sounds like something the various distro maintainers should obtain ASAP, at least if they expect customers who wish to run decently loaded HTTP servers with fewer worries.
You're probably mixing up "WindowMaker" -- a very specific window _manager_, ala fvwm, twm, mwm, blackbox -- with "GUI program designer" or something like that.
VB has naught to do with window _managing_, which involves placement of windows, title bars, widget sets, and so forth. WindowMaker has naught to do with writing programs, any more so than any other WM.
To answer your Q 'tho, well: get 'Learning Perl' and it's not that bad. For GUI, consider Java or Tcl/Tk if you must have portability, and there are probably other candidates, too. C w/ XForms, 'tho, is *fast* for programs that require X...
Hmmm. How many mobos would Asus need to sell to make a new design profitable? ...and can AMD guarantee a large enough production run?
The one thing about Intel is that it _will_ sell. Apparently, they can toss in processor ID weirdness and a mostly-incremental set of improvements and practically create another brand, and it will _still_ sell like mad. They could probably paint their CPUs orange, top 'em off with purple heat sinks shaped like drunk, lustful goats, and _still_ sell 'em -- in vast quantities, even if leading reviewers publicly question Grove's sanity.
AMD's ability to pump out a large volume has been repeatedly questioned. Not its ability to generate demand -- there's enough of that at the techno-geek level, most likely -- but to fulfill it...
Aye. As long as you've got all the needed libraries with no conflicts (I *think* that dyn-linking an X11R6 application with libc6 and X11R6 libraries dyn-linked with libc5 can do evil things, but it's been a while...), which is increasingly rare, you're fine.
Apps be distributed with the various libraries, or statically linked. This is the same issue w/ Windows, only somewhat better (in that DLLs tend not to have version numbers in their names, IIRC, and thus this makes it hard to have multiple versions...).
I've used Visual C++ on an NT box, but in general it makes me a tad less happy then I am devving on a Solaris or Linux box w/ XEmacs, make, and gdb... but that's probably largely because of the work I normally do.
I loathe the project configuration/maintenance in VC++: give me Makefiles (and not the strange ones that VC++ exports, which in my experience A) usually don't work w/ nmake.exe, and B) are not exactly designed for readability or re-usability).
The other major beef is more with MSDN than the IDE: the documentation seems a lot less... functional than the man pages that I'm used to for Solaris and Linux. I'm used to seeing, for instance, mentions of which header files and libraries to reference for a particular function; any implementation bugs regarding behavior versus existing standards; and so forth -- and I'm used to man-page-style documents being reasonably *up to date*.
That's not to say VC++ has no redeeming features; it does. But if you intend to port to other OSes, or don't have (paper) Win32 programming references around, or so forth -- you might be better off with a different package.
Hmmm. What's so difficult about looking at the technical specs section of a monitor manual? Reading? Searching the table of contents? The actual typing?
C'mon.
Heh. Why use a microkernel? Go ask Torvalds, or read his response to a fairly pro-monolithic-kernel chap in the Letters section of a recent "Communications of the ACM" (don't have the issue w/ me; it's the one that has a special section on lifelike graphics). It's a pretty interesting read.
That's basically the conclusion that I drew: to stop interception, you must be able to use a format that nobody can read unauthorized, but current destinations (audio devices) weren't exactly designed that way.
How do you send a "secret" message to somebody who does not know any encryption algorithms and cannot keep *any* secrets (including those in his head)? This is an analogous problem: the audio signal is *everything*, but also happens to be necessary -- and it's basically cleartext from a pirate's POV.
As somebody else noted, require special hardware on the receiving end, as hardware reverse-engineering would make this a lot more difficult, most likely.
then it can be intercepted, it would seem. So, break one of the assumptions: require audio devices that can handle the protected data, or block all forms of interception (which might be tricky if one includes attacks on the physical connection?). Ugh.
Well, perhaps there's a more elegant way, but it's not as obvious as a giant mutant glowing crow in a snowfield.
Hmmmmm. I'm not a moderator, but perhaps there should be a "DIE THREAD DIE" option. Or some other restriction on obvious cascades, which seem to be either
a) two camps with immutable and opposing opinions battling each other on an issue, or
b) (today at least) pointless content-free cascades.
Now I'll have a few seconds of warning in which to scream before a horde of those Xenomorph aliens and face-huggers comes a-runnin'. :)
"They're inside the perimeter!"
La la laaaauuuuuuuuugh!
Hmmmm. I see mentions of subscriptions on SuSE's (English) site, but their ordering page doesn't seem to list subscriptions as an option -- just their assorted one-time-purchase packages. Odd. Perhaps that's just lagging a bit?
One example is at LinuxMall.
$ insmod toaster.o /dev/ts0 /mnt/toaster
/dev/ts0: Error: toaster on fire. Suggest emergency eject. /dev/ts0
/dev/ts0: Toaster in use, ejection lever locked. /mnt/toaster /dev/ts0 not responding, possibly on fire. /dev/ts0 is still mounted.
ToastMaster detected. Installing toast-1.1 module.
$ mount
$ eject
$ umount
Error:
$ rmmod toaster.o
Cannot unload module.
Aaaaaaaiiiiiiiiiiieeeeeeeeeeeee! :)
-- Excerpted from Digital Screams: Why I Hate the Transistor, by Joh'omme Diwaand, pub. 2137 by Retro Press.