Here's some real info for those of you who couldn't figure out what was true in and among all of the conflicting info posted here.
1. uwm did have icons of a sort. You could iconify a window, and it would place a small text box where the window had been with the name of the window in it. The great part was you could edit the name on the fly!
2. No decorations, a window had what it was born with. All actions performed on a window were done with key-mouse combinations.
3. uwm was not the first X window manager, that would be wm.
There are many others, but those are the ones I can think of.
Hope this helps folks out. I started using X around the tail end of X10. uwm was the only window manager that shipped with X as far as I could tell, but it was enough for me at the time.
Now I wonder how I ever got along without at least fvwm2's "sloppy focus" (which everyone supports in one way or another, now).
At the Department of Transportation, I worked on the ETMS (Enhanced (air) Traffic Managment System), which is the system that the national airspace controlers use to figure out, e.g., if there's going to be congestion over Chicago today and re-route or delay planes to avoid it.
The system is something that has evolved over the years, and could have benefitted from a clean re-implimentation using modern tools and protocols. The problem was that they would have had to spin off a VERY large project to do so, and failures in the real-time traffic management program had made such programs political footballs.
I proposed a solution: Open Source.
Take all of the code, clean out anything that could a) be used to determine how critical systems (e.g. real-time air traffic control) worked, and there weren't many of them or b) could indicate how the security of the current system functions. Then establish a panel of 2-4 people who act as gatekeepers for the source. They release the source to the world and organize 5-10 projects around replacing the code from the ground up. No one ever took me up on it, but I think it would have resulted in one of the best government systems ever designed. Certainly your average OSS project is much better designed than any government system I've ever seen.
The gatekeepers would be responsible for code reviews on all incoming check-ins, and no one but the gatekeepers would have write-permission to the original source tree (though, you'd probably do something like sourceforge for the external developers to use as a sandbox). It's really no less secure than hiring random contractors to work on the code.
Comparison with Applied Cryptography?
on
Code Breaking
·
· Score: 3
I'm currently engrosed in reading my on a winding path through Applied Cryptography. It's not exactly a history of encryption, but I think it covers a lot of the bases that you mention from the point of view of teaching you how to DO crypto, not just LOOKING at it.
Sounds like most your time would be well served checking out Applied Cryptography first and settling on the more historical books if you find that A.P.'s not your bag of tea (and it won't be most people's).
That said, everyone who programs should read the first couple of chapters, the section on politics and the afterward of A.P. Skimming the protocols is also a good idea even if you just have to touch a high-level protocol (like SSL) from a distance.
A.P. is already getting me excited about some new projects. I started trying to write a pure-Perl implimentation of Diffie-Helman, and found myself side-tracked on trying to do efficient prime-generation. I've no decided to fork off into creating a repository of example perl programs of which my "mkprime" will be the first. Perl is a vastly under-used teaching language with the best features of C++ (for hiding some of the more complex details like arbitrary-precision math), LISP (for abstraction and algorithmic clarity) and C (for access to low-level details).
Being bisexual, and having a lot of friends who are gay or bisexual in this industry, my guess is this: when you're a teen and feeling alienated because football and other "turf war" games don't hold your attention, you find something else to do. For some, it's art. For some it's computers. So, you take this higher-than normal incidence of computer-using gay and bisexual men (gay women seem not to end up as enamoured with computers on the average) and throw them at the college sceene. Obviously they will not be selecting colleges for their sports program. Thus, the technical colleges get a surge in gay and bisexual interest.
It's only a guess, and I'm not sure that it explains the VERY unusual numbers, nor the fact that there appears to be more gay and bisexual clustering as you move up to the over-achivers in the field.
On a funny note, the best way I ever came out to someone (other than posting to Slashdot;-) was when a (straight) friend of mine and I were at dinner and he said, "well, Aaron, all the best sysadmins are fat, bearded and gay. So, I guess 2 out of 3 isn't bad." My reply was simply, "actually, that'd be 3."
I just waited for a moment as he realized what I'd just said, and we both laughed.
Very good point. I think that there are other good reasons to fork or have competing implimentations, but this is an excellent one.
We don't, however, have to go all the way back to the potato famine to see the damage that this sort of heterogeneous community has. Take a look at the damage that the infamous Morris Worm wreaked. It only knew how to exploit sendmail and either ftpd or fingerd on Suns and VAXes....
Looking at the responses in the article, this one was pretty much a summation of the (more verbose) others:
Security Admin (security@cyberlink.ch)
VERY harmful. This is screaming for a code-fork, for the same procedure that happend with SSH. If ISC doesn't back off, we're soon gonna have OpenBind.
So, to me it's pretty obvious that this move is going to produce a fork. Someone does mention dents which might provide another path entirely....
The operative question is: Is this a good thing? I think the track record (ssh/openssh, RSA/PGP/GPG, GIF/JPG/PNG, UNIX/BSD/Linux, etc, etc) shows that the answer will eventually be yes.
Some more than others. I've seen a number of Monty Hall type Dungeons and Dragons "dungeon crawls" where the goal is essentially to identify a "bad guy" who is weak enough to overpower and steal from. In this case it's a simple "he's got it; I take it" proposition.
In most modern games this *can* be replicated, but that's not as clearly the goal in the system.
Systems that come to mind are:
White Wolf's Mage: The Ascension. More so than any other game, this one tries to get the point to be the story and the characters and not the external conflict (though there's lots of that).
FASA's Shadowrun. Here's a system where the goal is to do your job and get paid.... There's just one catch: your job is to be a cyberpunk-cum-magic-ish mercenary, selling your services to the mega-corps for smuggling, theft, or whatever other corporate backstabbing the corps can't admit to in public.
Traveller (as well as GURPS Traveller, the GURPS-ized re-release of the system). A finer space merchant / mercenary / millitary game has not been devised as far as I can tell. Here, your goal often was related to trading or millitary actions depending on what sort of flavor campaign was being run.
I think the non-zero-sum nature of true RPGs is what will always stand between them and the computer-game faux-RPGs. Some games (e.g. Soul Reaver, Dungeon Keeper) try to be true to this RPG legacy, but there are very few.
Mason is pretty nice, I checked it out when it was just getting started. They've done a lot since then.
The high-level features are outlined on their site. Among them are: caching of HTML and data, use of the Perl debugger, and staging vs. production execution modes.
The CPAN repository is chock full of mind-bogglingly useful code, so it's nice to see more tools coming of age that allow you to take advantage of all of that through the Web. mod_perl is very nice, but for some tasks, it's just too cumbersome to work at that level. Mason nicely abstracts that away.
Milton -- Toy companies (followup: Dante/classic authors of religous fiction)
Picasso -- Artists (full circle) (followup: Rembrandt)
Pendragon -- Arthur (followup: Emperor/name for an absolute ruler)
Light -- Names Jesus called himself
Pirate -- Sea-based professions
Clancy -- MFK Fisher and Judith Clancy wrote "Not a Station But a Place" (followup: Lambert/Highlander)
Cougar -- Carnivorous mammals of North America (followup: Firebird/Car)
Have to admit, I'm partial to this last one, but that's just because I'd love the next ".0" release to be called "Phoenix";-)
Re:Microsoft can't do anything about free..
on
Linux Is Going Down
·
· Score: 2
Yes, Linux is not about money. However, if Microsoft can convince the market that Linux is not only not about money, but not a stable business, they will win back huge market-share from businesses that will not use a product that does not come from an identifiable, stable business.
Let's see if regular expressions improve this glop any, shall we?
"Based on the warnings from the developers and confusing messages from the company, it is clear the long-heralded Windows 2000 is a long way from being ready for business use," said Sherman. "The features in it are just the beginning, still raw technology."
But some Windows developers say that the newly released OS was just meant to be a beginning.
"Windows 2000 does lack serious system management tools, but there's a slew of new products coming down the ramp very soon that will bring Windows into even more enterprises," said someone....
Pardon if this answer is a slight bit distracted, I'm doing this and an OS install at the same time on two machines.
First of all, databsses can handle large amounts of arbitrary data, such as BLOBs or big chunks of text.
Try:
grep -i 'contact address'/var/lib/mysql/maildb
As it turns out, this is not as good an idea as it might sound....:-(
Secondly, I think you missed the fact that maintaining two separate data stores (database for headers, filesystem for message content) will certainly be more work than just using one or the other.
insert into headers values ('foo@bar.com', 'Re: This is just a test', '/var/spool/joeuser/inbox/13948')
Doesn't seem to hard to me. As far as consistency checking goes, you can ignore the on-disk text except for displaying the message. If you want to use the headers from the file to refresh the database in the event of coruption, fine, but it's not a big requirement.
Lastly, a filesystem and a database are both stored on a magnetic disk (for the most part). How is it easier to corrupt one than the other?
Any backups of my data that I keep are also stored in the same physical universe, but I don't use this as an excuse not to keep backups. Having the headers lying in a plain text file to sanity-check against can only help.
So, what do you do in th ereverse case from what you described? Suppose the portion of the filesystem containing the mail is lost. Can you rebuild those lost messages from the database which you would have store only the headers?
Generally when one assesses risk, one works with cost/benefit tradeoffs. What you propose is very costly in terms of database resources, whereas duplicating headers on disk is very cheap. This cost comes in terms of disk space, time used to duplicate the data (which in a very large system could be staggering for every message body), etc.
I think you will find that the benefits of storing headers twice will far outweigh the cost of having done so. I can't say the same for storing open-ended (in terms of size) message bodies in a relational database.
I say do one or the other, and then BACKUP OFTEN!
Nice idea, but we're talking about software design here, not system administration procedures. Clearly a sysadmin should be backing the data up, but to tell the user, "something looks odd here, go chase down a sysadmin and make him restore a backup," is a lot less friendly than, "I found some courupt headers in message 501719, correcting..."
Time for us to put up or shut up
on
Clever Girl Bess
·
· Score: 3
So, everyone's upset because bess is a no-win scenario for schools and it's competition is falling by the wayside.
Something tickles the back of my brain... large software that many are forced to use... sucks... manipulative marketing... Oh yeah, this is why we write open source software!
So, let's put up or shut up. Here's some specs, anyone up for implimentation and organization?
Target audience is broken into three: users (kids) organizations (schools, libraries, parents, etc) and BSPs (blocking service providers). In many cases, organizations will be BSPs, or will band into small groups to form one (e.g. school districts)
For the users, you provide several of the features for controling access in the OTHER direction (e.g. adblocker type features as well as some basic anonymity)
For the organizations, you provide sweeping, but simple controls over level of control. For example, you can turn logging on or off; you can select the classes of content to block; you can select the sources of blocking recomendations (there's a business model in there, even for Bess).... It's also important for the organizations to be able to delegate certain functions to users at their option. If, for example, the organization wants to block commercial sites by default, but allow users to turn them on, that should be OK.
For the BSPs, there's controls for who can use the service; what options are allowed and how much security is in place.
It should be administered entirely through a Web browser.
Users should be organized by class so that each class can have its own profile (e.g. "student under 13", "student over 13", "students in physiology class", "teachers", etc.). Profiles must be defined by the BSP or the organization, since no one-size-fits-all model will be complete.
Log file reporting must be at least up to basic usage reports
The system should be as platform neutral for organizations and users as possible
There should be a feedback mechanism so that users can feed back to organizations, and organizations to BSPs on incorrect blockage (an appeal system).
I'd start coding this myself, but I'm working on another mostly-open-source project in my free time, which I think a lot of people will like, and which might even end up being as socially relevant if I do it right.
Email messages are a specifically interesting topic. They're (for the most part) text, and tend to be larger than database fields want to be (on the order of 1+ kB each ranging all the way up to many megabytes in common practice).
This makes most mail messages poor choices for database storage (for example you want to be able to use "grep" on mail or compress in-place. Headers on the other hand are a major win in a database ("select messageid from headers where user = 'me' and date > yesterday and fromaddr = 'taco@slashdot.org'" should be fast even if I have tens of thousands of messages).
The easy solution is to keep the headers in the database, and then just keep maildirs with the original messages in the normal filesystem with the filenames in the database with the headers (something like message.headerid => headers.id and message.text is a path to the maildir entry for this message.
This combines the best of both worlds. This also means that while it's easy to corrupt your database with a single bug in your code, you can always re-build it from the on-disk messages.
Red Hat has been losing mindshare as of late, especially with the disaster that was Red Hat Linux 7.0
Why do people insist on repeating this mantra every time Red Hat comes out with a release (I started seeing this around 5.2). Every release of every OS has it's problems, but Red Hat 7.0 was probably the single most stable ".0" release I've seen from any vendor (including Microsoft, Sun, HP, etc).
The crypto is all there and as solid as can be. People whine about the GCC version, but even the GCC folks admit that there was no option that could have preserved any kind of compatibility with C++ binaries past-and-future. That's really a GCC problem, not a Red Hat problem. Red Hat just took the more obvious and therefore less politically acceptable route.
On the other hand, the integration of XFree86 4.0 and GNOME are beautiful as are many of the new features in up2date and the "preview" materials including a well-integrated 2.4 pre-release kernel.
All things considered, not a release for the front-end-production-server (mostly because there are so many new things that you should be taking a few months to figure out how to best deploy them), but then what ".0" release is (Solaris 2.0? 8.0? NT 4.0?) Seriously though, if you bought Red Hat 7.0, and you didn't like it, feel free to mail me the unregistered box, and I'll be happy to sell it off on ebay for my own benifit;-)
You're right, of course, but in the Western press (and I mean in the US primarily but most other countries as well with the exception of Canada) Akira was widely noted as having broken the mold in the sense that it was a commercial success. Fantasia did not do very well when it came out because it wasn't just for kids. Heavy Metal, Rock 'N Rule and other adult-subject western animations were always critisized and marginalized. Now, you see hollywood proclaiming that animation isn't just for kids anymore.
What changed? Basically it was the aftermath of Akira and several other movies from Japan (Ghost in the Shell, Mononoke Hime, etc) along with a few key movies in the West (most notably Iron Giant for having gotten crittical aclaim for it's adult themes in the face of zero expectations and effort on the part of the studio). The long-running success of the animation festivals has not hurt either.
I lay a lot of credit at the feet of Akira for waking the west up, not for inventing anything.
Is Linus killing Linux? Of course, as well he should. Linux could go in any ONE of thousands of directions. Linus has been pruning that probability tree for years now, and it shows. Linux could have gone the old-guard direction that the BSDs are maintaining. It could have become the bloated feature-pig that Windows is. It could have become an SMP-workhorse.
As it is Linus has taken pieces from each of these paths, but created something which is uniquely Linux. This is neither bad nor good, you have to judge it on its own merits. But, in the end, he has killed all of those could-have-been Linuxes. Some of those options would have benifited companies like Red Hat or SuSE more than what we have now, but if it's really too much of a burden to go down Linus' path, any one of these organizations can fork.
Hell, a fork can be maintained in a very sane fashion if you really try. We did it at KSR while the company existed (it died for internal reasons, but the OSF/1 development was going strong).
Linus has killed Linux hundreds of times, and here's to hoping he keeps doing so for decades to come!
Ain't It Cool News has slightly more news on their site. In general, such items tend to show up on AICN first (from which Slashdot has gotten more than one headline in the recent past).
I look forward to the re-release. There were a lot of things about Akira that might seem minor today, but were HUGE at the time. The two that come right to mind are:
Akira was the first animated movie to use computer-generated wireframes for all of their perspectives. This is why the obviously hand-drawn cityscapes have perfect perspective through long, and often complex pans.
This was the movie that convinced the world that animation != cartoons. The level of gore and adult plot you might actually have to think about stunned most non-Japaneese that saw it (I know I was one).
Gtk+ is one of the finest examples of software engineering I've ever seen. By extension, I'm including glib. Check out these documents for most of my reasoning:
So, the idea is that Enlightenment will be fast because the tons of junk that it does will be hardware-accelerated? Wouldn't it be logical, then to assume that all of the other window managers out there that are faster than Enlightenment (Sawfish, fvwm{,2}, twm, etc, etc,) will be even faster than Enlightenment given the same hardware acceleration?
Sawfish seems to me to be about 2-10 times faster (purely subjective) than Enlightenment. Can that gap be bridged by hardware? Of course, but it still doesn't make E efficient.
1. Unblock the mail
2. Follow the URL in the spamcop messages
3. Follow their instructions for a) noting that the problem has been addressed and b) sign up as an ISP and set your domains' abuse email addresses appropriately.
SpamCop sends email to you only when a) you have not registered a given URL or from-email address as being spam-free or b) someone with a spamcop account appeals your assertion. Since you've never bothered to feed back into the system, they keep pestering you (and rightly so).
I discovered spamcop as an "ISP" who got mail from them, and I joined because I love the way they deal with complaints (even those directed at me).
Check out SpamCop for a great way to help deal with spam. SpamCop is free to use, but you can also sign up and pay them money.
Really, paying them money is to support their work, but you also get a spam-free email forwarding service where yourname@spamcop.net gets forwarded to your favorite mail drop without any of the spam (which they do a very good job of filtering).
I'm using harmil@spamcop.net now as my primary "public" email address for things like slashdot and USENET, and it works pretty well.
Their spam reporting service is very cool. It tracks down the ISP of the spammer, submits the IP address of relays to ORBS, and also tracks any URLs in the spam body. Plus, ISPs who play ball with SpamCop can mark accounts as deleted and otherwise feed back into the system to reduce their request-load. Such things can be appealed by paying SpamCop users, but for the most part, ISPs are pretty good about it.
I just happen to be wearing my "Free the Berkeley 4.4!" shirt today from back when the AT&T vs. Berkeley suit was on everyone's mind, and we all wondered if we'd become wholy owned subsidiaries of the death star.
My other favorites:
The Madelbrot set USENIX shirt (Cincinatti, I think)
The OSI network model with "Financial" and "Political" added at the top of the stack with an arrow pointing to "Political" and a label: "You are here"
And for recent additions, I think "got root?" takes it.
Slashdot Mirror
Here's some real info for those of you who couldn't figure out what was true in and among all of the conflicting info posted here.
1. uwm did have icons of a sort. You could iconify a window, and it would place a small text box where the window had been with the name of the window in it. The great part was you could edit the name on the fly!
2. No decorations, a window had what it was born with. All actions performed on a window were done with key-mouse combinations.
3. uwm was not the first X window manager, that would be wm.
4. The timeline is something like this:
wm
uwm
twm
olwm, mwm
fvwm, gwm, tvwm, olvwm
fvwm2, windowmaker
enlightenment, icewm
sawmill (later sawfish)
There are many others, but those are the ones I can think of.
Hope this helps folks out. I started using X around the tail end of X10. uwm was the only window manager that shipped with X as far as I could tell, but it was enough for me at the time.
Now I wonder how I ever got along without at least fvwm2's "sloppy focus" (which everyone supports in one way or another, now).
At the Department of Transportation, I worked on the ETMS (Enhanced (air) Traffic Managment System), which is the system that the national airspace controlers use to figure out, e.g., if there's going to be congestion over Chicago today and re-route or delay planes to avoid it.
The system is something that has evolved over the years, and could have benefitted from a clean re-implimentation using modern tools and protocols. The problem was that they would have had to spin off a VERY large project to do so, and failures in the real-time traffic management program had made such programs political footballs.
I proposed a solution: Open Source.
Take all of the code, clean out anything that could a) be used to determine how critical systems (e.g. real-time air traffic control) worked, and there weren't many of them or b) could indicate how the security of the current system functions. Then establish a panel of 2-4 people who act as gatekeepers for the source. They release the source to the world and organize 5-10 projects around replacing the code from the ground up. No one ever took me up on it, but I think it would have resulted in one of the best government systems ever designed. Certainly your average OSS project is much better designed than any government system I've ever seen.
The gatekeepers would be responsible for code reviews on all incoming check-ins, and no one but the gatekeepers would have write-permission to the original source tree (though, you'd probably do something like sourceforge for the external developers to use as a sandbox). It's really no less secure than hiring random contractors to work on the code.
I'm currently engrosed in reading my on a winding path through Applied Cryptography. It's not exactly a history of encryption, but I think it covers a lot of the bases that you mention from the point of view of teaching you how to DO crypto, not just LOOKING at it.
Sounds like most your time would be well served checking out Applied Cryptography first and settling on the more historical books if you find that A.P.'s not your bag of tea (and it won't be most people's).
That said, everyone who programs should read the first couple of chapters, the section on politics and the afterward of A.P. Skimming the protocols is also a good idea even if you just have to touch a high-level protocol (like SSL) from a distance.
A.P. is already getting me excited about some new projects. I started trying to write a pure-Perl implimentation of Diffie-Helman, and found myself side-tracked on trying to do efficient prime-generation. I've no decided to fork off into creating a repository of example perl programs of which my "mkprime" will be the first. Perl is a vastly under-used teaching language with the best features of C++ (for hiding some of the more complex details like arbitrary-precision math), LISP (for abstraction and algorithmic clarity) and C (for access to low-level details).
Any suggestions on how to publicize such a site?
Being bisexual, and having a lot of friends who are gay or bisexual in this industry, my guess is this: when you're a teen and feeling alienated because football and other "turf war" games don't hold your attention, you find something else to do. For some, it's art. For some it's computers. So, you take this higher-than normal incidence of computer-using gay and bisexual men (gay women seem not to end up as enamoured with computers on the average) and throw them at the college sceene. Obviously they will not be selecting colleges for their sports program. Thus, the technical colleges get a surge in gay and bisexual interest.
;-) was when a (straight) friend of mine and I were at dinner and he said, "well, Aaron, all the best sysadmins are fat, bearded and gay. So, I guess 2 out of 3 isn't bad." My reply was simply, "actually, that'd be 3."
It's only a guess, and I'm not sure that it explains the VERY unusual numbers, nor the fact that there appears to be more gay and bisexual clustering as you move up to the over-achivers in the field.
On a funny note, the best way I ever came out to someone (other than posting to Slashdot
I just waited for a moment as he realized what I'd just said, and we both laughed.
Very good point. I think that there are other good reasons to fork or have competing implimentations, but this is an excellent one.
We don't, however, have to go all the way back to the potato famine to see the damage that this sort of heterogeneous community has. Take a look at the damage that the infamous Morris Worm wreaked. It only knew how to exploit sendmail and either ftpd or fingerd on Suns and VAXes....
The operative question is: Is this a good thing? I think the track record (ssh/openssh, RSA/PGP/GPG, GIF/JPG/PNG, UNIX/BSD/Linux, etc, etc) shows that the answer will eventually be yes.
Some more than others. I've seen a number of Monty Hall type Dungeons and Dragons "dungeon crawls" where the goal is essentially to identify a "bad guy" who is weak enough to overpower and steal from. In this case it's a simple "he's got it; I take it" proposition.
In most modern games this *can* be replicated, but that's not as clearly the goal in the system.
Systems that come to mind are:
White Wolf's Mage: The Ascension. More so than any other game, this one tries to get the point to be the story and the characters and not the external conflict (though there's lots of that).
FASA's Shadowrun. Here's a system where the goal is to do your job and get paid.... There's just one catch: your job is to be a cyberpunk-cum-magic-ish mercenary, selling your services to the mega-corps for smuggling, theft, or whatever other corporate backstabbing the corps can't admit to in public.
Traveller (as well as GURPS Traveller, the GURPS-ized re-release of the system). A finer space merchant / mercenary / millitary game has not been devised as far as I can tell. Here, your goal often was related to trading or millitary actions depending on what sort of flavor campaign was being run.
I think the non-zero-sum nature of true RPGs is what will always stand between them and the computer-game faux-RPGs. Some games (e.g. Soul Reaver, Dungeon Keeper) try to be true to this RPG legacy, but there are very few.
Mason is pretty nice, I checked it out when it was just getting started. They've done a lot since then.
The high-level features are outlined on their site. Among them are: caching of HTML and data, use of the Perl debugger, and staging vs. production execution modes.
The CPAN repository is chock full of mind-bogglingly useful code, so it's nice to see more tools coming of age that allow you to take advantage of all of that through the Web. mod_perl is very nice, but for some tasks, it's just too cumbersome to work at that level. Mason nicely abstracts that away.
Possible followups to Fisher:
;-)
Milton -- Toy companies (followup: Dante/classic authors of religous fiction)
Picasso -- Artists (full circle) (followup: Rembrandt)
Pendragon -- Arthur (followup: Emperor/name for an absolute ruler)
Light -- Names Jesus called himself
Pirate -- Sea-based professions
Clancy -- MFK Fisher and Judith Clancy wrote "Not a Station But a Place" (followup: Lambert/Highlander)
Cougar -- Carnivorous mammals of North America (followup: Firebird/Car)
Have to admit, I'm partial to this last one, but that's just because I'd love the next ".0" release to be called "Phoenix"
Yes, Linux is not about money. However, if Microsoft can convince the market that Linux is not only not about money, but not a stable business, they will win back huge market-share from businesses that will not use a product that does not come from an identifiable, stable business.
"Based on the warnings from the developers and confusing messages from the company, it is clear the long-heralded Windows 2000 is a long way from being ready for business use," said Sherman. "The features in it are just the beginning, still raw technology."
But some Windows developers say that the newly released OS was just meant to be a beginning.
"Windows 2000 does lack serious system management tools, but there's a slew of new products coming down the ramp very soon that will bring Windows into even more enterprises," said someone....
Doesn't seem to hard to me. As far as consistency checking goes, you can ignore the on-disk text except for displaying the message. If you want to use the headers from the file to refresh the database in the event of coruption, fine, but it's not a big requirement.
Any backups of my data that I keep are also stored in the same physical universe, but I don't use this as an excuse not to keep backups. Having the headers lying in a plain text file to sanity-check against can only help. Generally when one assesses risk, one works with cost/benefit tradeoffs. What you propose is very costly in terms of database resources, whereas duplicating headers on disk is very cheap. This cost comes in terms of disk space, time used to duplicate the data (which in a very large system could be staggering for every message body), etc.I think you will find that the benefits of storing headers twice will far outweigh the cost of having done so. I can't say the same for storing open-ended (in terms of size) message bodies in a relational database.
Nice idea, but we're talking about software design here, not system administration procedures. Clearly a sysadmin should be backing the data up, but to tell the user, "something looks odd here, go chase down a sysadmin and make him restore a backup," is a lot less friendly than, "I found some courupt headers in message 501719, correcting..."Something tickles the back of my brain... large software that many are forced to use... sucks... manipulative marketing... Oh yeah, this is why we write open source software!
So, let's put up or shut up. Here's some specs, anyone up for implimentation and organization?
I'd start coding this myself, but I'm working on another mostly-open-source project in my free time, which I think a lot of people will like, and which might even end up being as socially relevant if I do it right.
Email messages are a specifically interesting topic. They're (for the most part) text, and tend to be larger than database fields want to be (on the order of 1+ kB each ranging all the way up to many megabytes in common practice).
This makes most mail messages poor choices for database storage (for example you want to be able to use "grep" on mail or compress in-place. Headers on the other hand are a major win in a database ("select messageid from headers where user = 'me' and date > yesterday and fromaddr = 'taco@slashdot.org'" should be fast even if I have tens of thousands of messages).
The easy solution is to keep the headers in the database, and then just keep maildirs with the original messages in the normal filesystem with the filenames in the database with the headers (something like message.headerid => headers.id and message.text is a path to the maildir entry for this message.
This combines the best of both worlds. This also means that while it's easy to corrupt your database with a single bug in your code, you can always re-build it from the on-disk messages.
The crypto is all there and as solid as can be. People whine about the GCC version, but even the GCC folks admit that there was no option that could have preserved any kind of compatibility with C++ binaries past-and-future. That's really a GCC problem, not a Red Hat problem. Red Hat just took the more obvious and therefore less politically acceptable route.
On the other hand, the integration of XFree86 4.0 and GNOME are beautiful as are many of the new features in up2date and the "preview" materials including a well-integrated 2.4 pre-release kernel.
All things considered, not a release for the front-end-production-server (mostly because there are so many new things that you should be taking a few months to figure out how to best deploy them), but then what ".0" release is (Solaris 2.0? 8.0? NT 4.0?) Seriously though, if you bought Red Hat 7.0, and you didn't like it, feel free to mail me the unregistered box, and I'll be happy to sell it off on ebay for my own benifit ;-)
You're right, of course, but in the Western press (and I mean in the US primarily but most other countries as well with the exception of Canada) Akira was widely noted as having broken the mold in the sense that it was a commercial success. Fantasia did not do very well when it came out because it wasn't just for kids. Heavy Metal, Rock 'N Rule and other adult-subject western animations were always critisized and marginalized. Now, you see hollywood proclaiming that animation isn't just for kids anymore.
What changed? Basically it was the aftermath of Akira and several other movies from Japan (Ghost in the Shell, Mononoke Hime, etc) along with a few key movies in the West (most notably Iron Giant for having gotten crittical aclaim for it's adult themes in the face of zero expectations and effort on the part of the studio). The long-running success of the animation festivals has not hurt either.
I lay a lot of credit at the feet of Akira for waking the west up, not for inventing anything.
Is Linus killing Linux? Of course, as well he should. Linux could go in any ONE of thousands of directions. Linus has been pruning that probability tree for years now, and it shows. Linux could have gone the old-guard direction that the BSDs are maintaining. It could have become the bloated feature-pig that Windows is. It could have become an SMP-workhorse.
As it is Linus has taken pieces from each of these paths, but created something which is uniquely Linux. This is neither bad nor good, you have to judge it on its own merits. But, in the end, he has killed all of those could-have-been Linuxes. Some of those options would have benifited companies like Red Hat or SuSE more than what we have now, but if it's really too much of a burden to go down Linus' path, any one of these organizations can fork.
Hell, a fork can be maintained in a very sane fashion if you really try. We did it at KSR while the company existed (it died for internal reasons, but the OSF/1 development was going strong).
Linus has killed Linux hundreds of times, and here's to hoping he keeps doing so for decades to come!
I look forward to the re-release. There were a lot of things about Akira that might seem minor today, but were HUGE at the time. The two that come right to mind are:
Check out This insane patent which is referenced in the press release. I almost burst a gut!
Gtk+ is one of the finest examples of software engineering I've ever seen. By extension, I'm including glib. Check out these documents for most of my reasoning:
The Gtk+ FAQ
The gobject reference
Tic Tac Toe in Gtk+
Your sig includes the line:
/etc/fstab dos
/etc/fstab
grep
That I'm curious, why do you have a file called "dos", and why does it refer to your fstab? Or did you mean,
grep dos
So, the idea is that Enlightenment will be fast because the tons of junk that it does will be hardware-accelerated? Wouldn't it be logical, then to assume that all of the other window managers out there that are faster than Enlightenment (Sawfish, fvwm{,2}, twm, etc, etc,) will be even faster than Enlightenment given the same hardware acceleration?
Sawfish seems to me to be about 2-10 times faster (purely subjective) than Enlightenment. Can that gap be bridged by hardware? Of course, but it still doesn't make E efficient.
You've failed to use spamcop.
Here's what you want to do:
1. Unblock the mail
2. Follow the URL in the spamcop messages
3. Follow their instructions for a) noting that the problem has been addressed and b) sign up as an ISP and set your domains' abuse email addresses appropriately.
SpamCop sends email to you only when a) you have not registered a given URL or from-email address as being spam-free or b) someone with a spamcop account appeals your assertion. Since you've never bothered to feed back into the system, they keep pestering you (and rightly so).
I discovered spamcop as an "ISP" who got mail from them, and I joined because I love the way they deal with complaints (even those directed at me).
Check out SpamCop for a great way to help deal with spam. SpamCop is free to use, but you can also sign up and pay them money.
Really, paying them money is to support their work, but you also get a spam-free email forwarding service where yourname@spamcop.net gets forwarded to your favorite mail drop without any of the spam (which they do a very good job of filtering).
I'm using harmil@spamcop.net now as my primary "public" email address for things like slashdot and USENET, and it works pretty well.
Their spam reporting service is very cool. It tracks down the ISP of the spammer, submits the IP address of relays to ORBS, and also tracks any URLs in the spam body. Plus, ISPs who play ball with SpamCop can mark accounts as deleted and otherwise feed back into the system to reduce their request-load. Such things can be appealed by paying SpamCop users, but for the most part, ISPs are pretty good about it.
For the record, I'm just a customer.
I just happen to be wearing my "Free the Berkeley 4.4!" shirt today from back when the AT&T vs. Berkeley suit was on everyone's mind, and we all wondered if we'd become wholy owned subsidiaries of the death star.
My other favorites:
The Madelbrot set USENIX shirt (Cincinatti, I think)
The OSI network model with "Financial" and "Political" added at the top of the stack with an arrow pointing to "Political" and a label: "You are here"
And for recent additions, I think "got root?" takes it.