Enhydra is an open-source Java application server. It provides a graphical database design tool, a site management console, load balancing, database connection pooling, and session management which automatically chooses between using cookies and url rewriting. Plus it's free.:)
It also has a great tool called XMLC which IMHO is a far better solution for embedding dynamic content in HTML than ASP or JSP. It lets you create dynamic HTML pages which a graphic artist can then edit in Dreamweaver or whatever tools they want to use - they can change the look and feel of the site without bothering the programmers, and there are no non-standard tags for their tools to choke on.
The students will still have to do needs-analysis, design the database structure (if necessary), design the interface, write the servlets, integrate with existing legacy systems, integrate with any credit-card processing software they might need, etc... Enhydra just provides a fantastic set of tools to let their applications include advanced features which they might not otherwise have the time or inclination to implement. It also provides a supported platform that the businesses can grow on, add functions to, and scale up if their on-line requirements increase in the future... and you know they will.;)
this is precisely because he's wrong to claim there are moral absolutes, ethical prinicples which are culturally independent. There aren't. Ethical views are at least to some extent culturally determined SNIP
I strongly disagree with the first sentences, and wholeheartedly agree with the last one. Ethical views are to a great extent culturally determined, but that does not mean that there no universal standards of right and wrong. They're called instincts. Even lower animals shelter and protect their young.
You instinctively know when something's right or wrong, and we humans have developed a (sometimes very flawed) system of logically extending these strong feelings into abstract notions. The 'moral vacuum' philosophy has become popular recently, and you can logically argue that instincts are irrelevant, man has evolved beyond them, blah blah. Yeah, right. Look into your heart. Those 'instincts' determine what we value most highly, determine what we consider 'right' and 'wrong', and most importantly, they make life worth living.
I agree with Mr. Meyer on a few points, but overall his article was a long, long rant with few solid supporting arguments. Sure, Stallman is an extremist with some arguably wrong viewpoints, but his ideals are today bringing technology to the world's poorest countries. I think he's far too busy developing quality software to be wasting time with "character assassination performed on commercial software developers" And the whole gun control thing was completely irrelevant. Mr. Meyer comes off sounding shrill and irrational.
It sounds like this is something you want to implement in the next year or two, so I'll assume that you don't want to re-invent the wheel by writing your own database engine.
Use PostgreSQL or some other tested, reliable database. Put the data on mirrored & striped disks - RAID 0+1. Believe me, the extra dough is worth it, and performance will be fine.
Read up on data design. Too many people assume that once the database software is in place, laying out the database structure is the easy part. Then, after the application is in place for a while, problems start cropping up, and it can become a real maintenance nightmare. This most often happens with applications written in Access or Excel/VBA by well-meaning accountant types, but can happen to anyone. A database with incomplete and/or inaccurate data is worthless.
I recommend "Database Design for Mere Mortals", by Mike Hernandez, it offers a nice, practical methodology for data design, and is very clearly written.
I wasn't saying Linux has the same maintenance cost as Solaris. For those of you who want the quick rant, all I was saying was, "Look, using Microsoft's own study, Linux has lower TCO than NT, **EVEN IF** we assume Linux has the same maintenance costs as Solaris!!"
The assumption was necessary for the exercise, since I don't have any data to contradict Microsoft's paid "researchers". It was also partly intended to show how the numbers can be turned around to mean different things. For instance, why do you suppose they compared the price of 38 Solaris servers with the price of 30 NT servers?
My "claim" was that none of these studies are anywhere near objective. That's the only point I needed to make. ______________________________________________ Auditing and dentistry are excellent career choices for people who don't
People just haven't realized yet that none of these "objective" tech review groups have any reason to be truly objective. They claim objectivity, but there is no one holding them to it. The average PHB doesn't investigate the findings, and the money goes to whoever can consistently provide the best combination of perceived integrity and customer butt-kissing.
For example, in Microsoft's Linux Myths page, one of the key points is lower TCO. The study they quote to back this up was paid for by Microsoft and Compaq, and is seriously funny if you actually add up the numbers. The study is actually a comparison of NT TCO with Solaris/SPARC TCO, so since they use this "study" in an argument against Linux, I thought it would be appropriate to look at the numbers as they would appear in a Linux environment.
First of all, on every line, they compare the TCO of 30 NT servers to 38 Unix servers. Why? They don't say. In the absence of convincing evidence that 30 NT servers will do the job of 38 Unix servers, let's make this a server-to-server comparison and use numbers for 30 Unix servers instead. Let's assume the hardware costs are the same between NT and Linux, since they will both run on Intel hardware. That saves us big bucks over the Sparc hardware quoted in the study.
Now, for additional software, they include databases, development tools, apps, and utilities. Top quality Linux apps, dev tools, and utilities are free, but I could see paying money for Oracle or something on big boxes (no disrespect to MySql intended), so we'll include their figures for database expenditures, minus 21% to account for 30 servers instead of 38.
The initial purchase price and application prices I list below are double what are shown on the "TCO Summary" table on their web page. For some reason their summary figures are exactly half of the totals in their detail reports, and I couldn't determine why, so I went with the detail report. All the other summary figures on their page match their detail reports exactly.
All the support etc... stuff will probably be about the same between Solaris and Linux, so I just took those numbers right off their page.
So, even in their own study, Microsoft couldn't beat Linux in TCO. What Microsoft paid for was for BRG research to arrange the data in a way that complimented NT, even if they had to "fudge" the data. Imagine what the numbers would look like in a study the FSF paid for!
______________________________________________ Auditing and dentistry are excellent career choices for people who don't
As far as I can tell, the most significant difference is that you can't spool to CUPS from any existing lpd. According to the CUPS docs, it listens to the IPP port and can be configured to listen at the HTTP port, but...???? I could be wrong, but it looks like if you want to spool to a CUPS server from any other UNIX box, you have to install CUPS on that box.
That's not going to happen where I work.
Since many of the commercial UNIX implementations have added their own extra "features" to the ummm...was it RFC1170?... LPR protocol, writing a good full-featured lpd that works well with all of them can be tricky, and I think the CUPS folks decided not to bother. From what I've read, LPRng is the best "universal" implementation.
Again, please take with a grain of salt, all I know about CUPS is what I got from reading the Administration guide. If anyone knows differently, please correct me!!!!!
Auditing and dentistry are excellent career choices for people who don't
It's not Linux, its most likely the print filter you're using that was causing you headaches. You just gave up on using Linux too quickly. Other distributions such as Debian and Caldera install LPRng by default.
Use LPRng rather than Redhat's stock lpd - for a serious print server you want all the extra features LPRng offers, such as being able to send postscript directly to port 9100 on the Jetdirect card, web management (with lpinfo), the ability to have one printcap file that works across all print servers, support for installation on most commercial Unices, etc... LPRng comes on the Applications CD that comes with RedHat, is GPL'd for non-commercial uses, and is free for commercial implementations, but I'm unsure of other restrictions besides price for commercial uses. Anyway, if the license was good enough for the Debian group, it's good enough for me.
Well, CUPS is pretty useless to me, since it apparently does not accept LPD connections from other print servers. I dug through all their documentation, and as far as I can tell, if this is supported, it's not documented. And, CUPS uninstalls your stock lp* during installation. So, unless I install CUPS on every *NIX box on my network, I can't print to CUPS from *NIX?
According to the Admin docs, CUPS only listens on the IPP port, and you can set it up to listen to the HTTP port. No mention of LPD, other than to say it can SEND jobs via LPD.
By the way, I am near the end of a (so far) extremely successful implementation of LPRng in a commercial environment. I had a brief correspondence with Patrick Powell, author and current maintainer of LPRng, and he has assured me that I can use LPRng with no charge in a commercial environment, but I haven't seen the actual license... all it says in the README is "Released under the GPL for use in non-commercial environments," which is pretty vague for my taste.
Auditing and dentistry are excellent career choices for people who don't
The reality is that MicroSoft LIE when it gets broken (or broken into) - the pitiful "weather" excuse being clearly visible on globally accessible weather-radar images for the farce that it is
Well, as a resident of the Seattle area I can attest to the unusual electrical storms we were having last week, but that doesn't explain all their downtime. _______________________________________ ________________________________
And get their SIMS product, too, it's pretty well optimised for the high end.
According to what's currently coming down the grapevine, in the upcoming Sun/Netscape 'partnership' SIMS and Netscape's MTA products will be merged into one product, and sun will be replacing the buggy SIMS message store with the corresponding Messenger component. Hopefully this will make for one nice, solid server. _________________________________________ ______________________________
For 25000 users you want HIGH reliability. Any admin who put that many users on just one box has some headaches in store. __________________________________________ _____________________________
I think there are enough horror stories about Exchange bouncing around that I don't need to comment on that. I recommend Netscape Messenger. It's designed to work with high-availability clustering solutions, has built-in IMAP capability, and is packaged with Netscape Directory Server which means all your user account data is in accessible through LDAP, and can be shared with other applications. You would definitely want to run it over Solaris w/Veritas or some other nice, reliable OS with a journaling filesystem.
Sun has its own messaging product, SIMS, which offers all of the same features. Don't do it. Half the bugs I have reported have been responded to sort of as follows: "Well, it just does that. I think they're going to release a patch for that as soon as they figure out why." I have not been impressed with the SIMS support team. ___________________________________________ ____________________________
But we have been having power problems, the jokers who wired our new server room screwed up big time. We've also lost 4 power modules in an APC Symmetra that was powering it. Can't fault Sun. ____________________________________________ ___________________________
According to crayz, we've all been scanning the same 115 work units over and over again. That's a grand total of about 30 megs of data. Seti's own computers should be able to handle that. What a pisser. Unfortunately I don't have a record of the work units my WinDoze boxes have run, but here are the ones my home Linux box has done. Notice that there are several duplicates. Do these look familiar to anyone else?
Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency: 1.418886717 GHz Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency: 1.420126951 GHz Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency: 1.420126951 GHz Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency: 1.420185545 GHz Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency: 1.420439451 GHz Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency: 1.420683592 GHz Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency: 1.420869139 GHz Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency: 1.420976560 GHz Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency: 1.420976560 GHz Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency: 1.420976560 GHz Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency: 1.421142576 GHz Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency: 1.421142576 GHz Sky coordinates: 13.0 R.A., 12.7 Dec Recorded on: Fri Jan 8 00:51:05 1999 Base Frequency: 1.419101560 GHz Sky coordinates: 13.0 R.A., 12.7 Dec Recorded on: Fri Jan 8 00:51:05 1999 Base Frequency: 1.419101560 GHz Sky coordinates: 13.0 R.A., 12.8 Dec Recorded on: Fri Jan 8 00:49:44 1999 Base Frequency: 1.419404295 GHz Sky coordinates: 13.1 R.A., 12.2 Dec Recorded on: Fri Jan 8 01:05:49 1999 Base Frequency: 1.419960936 GHz _____________________________________________ __________________________
So what do we expect to find? Based on bandwidth usage on the Internet, my guess is that our first contact, once decoded, will turn out to be alien porno. __________________________________________ _____________________________
On Linux whichever had the higher priority ('niceness') would win and probably mostly squash the other one. You could set the priority the same, then they would compete for CPU time, but why crunch keys when you could be looking for ET's TV signals?
Hmm, based on bandwidth usage on the Internet, I wonder what the probability is that our first contact will be picking up alien porno... ________________________________________ _______________________________
Last time I looked, there were over 525,000 people registered, and the average processing time for one 250,000 byte work unit is currently just under 40 hours (My P350 does one in about 20 hours). So, that works out to about.00625 MB per computer per hour. With 525,000 people registered, that's at least 3281.25 MB per hour being processed, or 78750 GB per day. More than twice as much data as they're pulling in from Arecibo. Of course, that's assuming all the computers are working on it 24 hours per day, but I'll bet most every computer is putting in at least 12 hours of CPU time on this per day. So, we're already processing the data faster than they are taking it in. _____________________________________________ __________________________
Because everyone is analyzing the same 30 megabytes of data. One Linux box I am running it on has processed 16 work units. FIVE of them have been duplicates.
It is a waste of bandwidth - with 525,000 users it's somewhere between 35 and 70 GB per day assuming their statistics are correct (average 40 hours per work unit), and that each of the 525,000 users put in between 12-24 hours of CPU time per day. I was putting in about 60 hours total CPU time per day until just now, and averaging about 20 hours per work unit, so these numbers are probably low.
I have stopped running the client on all my machines until I hear that they are sending out new data. Once I hear there's new data, I'll start up the clients again. I love the idea, I just wish they hadn't been so premature about releasing the clients and announcing the project everywhere. _____________________________________ __________________________________
Jane Dark's feature, "Suffer the (White, Middle-Class) Children" was a poorly researched, irresponsible piece obviously designed to be inflammatory. She basically calls the oppressed high-school geeks whiners. She says it is a white problem, and implies that these white folks don't know what suffering is.
Well, she may be correct on the first count - in 1995 over 90% of suicides were by white folks. I think those people knew what suffering is. "Jane Dark" says she had her share of jerks to deal with in high school, and sure it sucked, but she was able to deal with it. I don't think she realizes the extent of the abuse some of these kids endure. Many of them are told daily that they are worthless human beings, a waste of oxygen. Maybe this doesn't happen as much in predominantly minority high schools. Whether or not that is true, the suffering exists, and it is extreme enough to drive these kids to kill themselves and each other. There is no point or purpose in making this a racial issue except to be inflammatory.
Jane, you have succeeded in provoking a response from me. I'm just sorry you had to belittle the suffering of thousands of teenagers to do it. _____________________________________________ __________________________
Slashdot Mirror
They're giving us lawyers a bad name!
Who, DeBrand's volunteer or PanIP's lawyer?
Enhydra is an open-source Java application server. It provides a graphical database design tool, a site management console, load balancing, database connection pooling, and session management which automatically chooses between using cookies and url rewriting. Plus it's free. :)
;)
It also has a great tool called XMLC which IMHO is a far better solution for embedding dynamic content in HTML than ASP or JSP. It lets you create dynamic HTML pages which a graphic artist can then edit in Dreamweaver or whatever tools they want to use - they can change the look and feel of the site without bothering the programmers, and there are no non-standard tags for their tools to choke on.
The students will still have to do needs-analysis, design the database structure (if necessary), design the interface, write the servlets, integrate with existing legacy systems, integrate with any credit-card processing software they might need, etc... Enhydra just provides a fantastic set of tools to let their applications include advanced features which they might not otherwise have the time or inclination to implement. It also provides a supported platform that the businesses can grow on, add functions to, and scale up if their on-line requirements increase in the future... and you know they will.
http://www.enhydra.org
this is precisely because he's wrong to claim there are moral absolutes, ethical prinicples which are culturally independent. There aren't. Ethical views are at least to some extent culturally determined
SNIP
I strongly disagree with the first sentences, and wholeheartedly agree with the last one. Ethical views are to a great extent culturally determined, but that does not mean that there no universal standards of right and wrong. They're called instincts. Even lower animals shelter and protect their young.
You instinctively know when something's right or wrong, and we humans have developed a (sometimes very flawed) system of logically extending these strong feelings into abstract notions. The 'moral vacuum' philosophy has become popular recently, and you can logically argue that instincts are irrelevant, man has evolved beyond them, blah blah. Yeah, right. Look into your heart. Those 'instincts' determine what we value most highly, determine what we consider 'right' and 'wrong', and most importantly, they make life worth living.
I agree with Mr. Meyer on a few points, but overall his article was a long, long rant with few solid supporting arguments. Sure, Stallman is an extremist with some arguably wrong viewpoints, but his ideals are today bringing technology to the world's poorest countries. I think he's far too busy developing quality software to be wasting time with "character assassination performed on commercial software developers" And the whole gun control thing was completely irrelevant. Mr. Meyer comes off sounding shrill and irrational.
It sounds like this is something you want to implement in the next year or two, so I'll assume that you don't want to re-invent the wheel by writing your own database engine.
Use PostgreSQL or some other tested, reliable database. Put the data on mirrored & striped disks - RAID 0+1. Believe me, the extra dough is worth it, and performance will be fine.
Read up on data design. Too many people assume that once the database software is in place, laying out the database structure is the easy part. Then, after the application is in place for a while, problems start cropping up, and it can become a real maintenance nightmare. This most often happens with applications written in Access or Excel/VBA by well-meaning accountant types, but can happen to anyone. A database with incomplete and/or inaccurate data is worthless.
I recommend "Database Design for Mere Mortals", by Mike Hernandez, it offers a nice, practical methodology for data design, and is very clearly written.
I wasn't saying Linux has the same maintenance cost as Solaris. For those of you who want the quick rant, all I was saying was, "Look, using Microsoft's own study, Linux has lower TCO than NT, **EVEN IF** we assume Linux has the same maintenance costs as Solaris!!"
The assumption was necessary for the exercise, since I don't have any data to contradict Microsoft's paid "researchers". It was also partly intended to show how the numbers can be turned around to mean different things. For instance, why do you suppose they compared the price of 38 Solaris servers with the price of 30 NT servers?
My "claim" was that none of these studies are anywhere near objective. That's the only point I needed to make.
______________________________________________
Auditing and dentistry are excellent career choices for people who don't
People just haven't realized yet that none of these "objective" tech review groups have any reason to be truly objective. They claim objectivity, but there is no one holding them to it. The average PHB doesn't investigate the findings, and the money goes to whoever can consistently provide the best combination of perceived integrity and customer butt-kissing.
. ..........Linux
For example, in Microsoft's Linux Myths page, one of the key points is lower TCO. The study they quote to back this up was paid for by Microsoft and Compaq, and is seriously funny if you actually add up the numbers. The study is actually a comparison of NT TCO with Solaris/SPARC TCO, so since they use this "study" in an argument against Linux, I thought it would be appropriate to look at the numbers as they would appear in a Linux environment.
First of all, on every line, they compare the TCO of 30 NT servers to 38 Unix servers. Why? They don't say. In the absence of convincing evidence that 30 NT servers will do the job of 38 Unix servers, let's make this a server-to-server comparison and use numbers for 30 Unix servers instead. Let's assume the hardware costs are the same between NT and Linux, since they will both run on Intel hardware. That saves us big bucks over the Sparc hardware quoted in the study.
Now, for additional software, they include databases, development tools, apps, and utilities. Top quality Linux apps, dev tools, and utilities are free, but I could see paying money for Oracle or something on big boxes (no disrespect to MySql intended), so we'll include their figures for database expenditures, minus 21% to account for 30 servers instead of 38.
The initial purchase price and application prices I list below are double what are shown on the "TCO Summary" table on their web page. For some reason their summary figures are exactly half of the totals in their detail reports, and I couldn't determine why, so I went with the detail report. All the other summary figures on their page match their detail reports exactly.
All the support etc... stuff will probably be about the same between Solaris and Linux, so I just took those numbers right off their page.
............................................NT.
Hardware+OS.....$684,980......$530,400
Applications...........$49,510.........$32,307
Support, etc..........$867,740...$1,035,496
TCO/Year.........$1,602,230...$1,598,203
So, even in their own study, Microsoft couldn't beat Linux in TCO. What Microsoft paid for was for BRG research to arrange the data in a way that complimented NT, even if they had to "fudge" the data. Imagine what the numbers would look like in a study the FSF paid for!
______________________________________________
Auditing and dentistry are excellent career choices for people who don't
As far as I can tell, the most significant difference is that you can't spool to CUPS from any existing lpd. According to the CUPS docs, it listens to the IPP port and can be configured to listen at the HTTP port, but...???? I could be wrong, but it looks like if you want to spool to a CUPS server from any other UNIX box, you have to install CUPS on that box.
That's not going to happen where I work.
Since many of the commercial UNIX implementations have added their own extra "features" to the ummm...was it RFC1170?... LPR protocol, writing a good full-featured lpd that works well with all of them can be tricky, and I think the CUPS folks decided not to bother. From what I've read, LPRng is the best "universal" implementation.
Again, please take with a grain of salt, all I know about CUPS is what I got from reading the Administration guide. If anyone knows differently, please correct me!!!!!
Auditing and dentistry are excellent career choices for people who don't
Yeah, if you can use a printer that supports Postscript directly rather than having to use Ghostscript, it will work better.
Auditing and dentistry are excellent career choices for people who don't
It's not Linux, its most likely the print filter you're using that was causing you headaches. You just gave up on using Linux too quickly. Other distributions such as Debian and Caldera install LPRng by default.
_ _______________
Use LPRng rather than Redhat's stock lpd - for a serious print server you want all the extra features LPRng offers, such as being able to send postscript directly to port 9100 on the Jetdirect card, web management (with lpinfo), the ability to have one printcap file that works across all print servers, support for installation on most commercial Unices, etc... LPRng comes on the Applications CD that comes with RedHat, is GPL'd for non-commercial uses, and is free for commercial implementations, but I'm unsure of other restrictions besides price for commercial uses. Anyway, if the license was good enough for the Debian group, it's good enough for me.
_______________________________________________
Auditing and dentistry are excellent career choices for people who don't
Well, CUPS is pretty useless to me, since it apparently does not accept LPD connections from other print servers. I dug through all their documentation, and as far as I can tell, if this is supported, it's not documented. And, CUPS uninstalls your stock lp* during installation. So, unless I install CUPS on every *NIX box on my network, I can't print to CUPS from *NIX?
According to the Admin docs, CUPS only listens on the IPP port, and you can set it up to listen to the HTTP port. No mention of LPD, other than to say it can SEND jobs via LPD.
By the way, I am near the end of a (so far) extremely successful implementation of LPRng in a commercial environment. I had a brief correspondence with Patrick Powell, author and current maintainer of LPRng, and he has assured me that I can use LPRng with no charge in a commercial environment, but I haven't seen the actual license... all it says in the README is "Released under the GPL for use in non-commercial environments," which is pretty vague for my taste.
Auditing and dentistry are excellent career choices for people who don't
The reality is that MicroSoft LIE when it gets broken (or broken into) - the pitiful "weather" excuse being clearly visible on globally accessible weather-radar images for the farce that it is
_ ________________________________
Well, as a resident of the Seattle area I can attest to the unusual electrical storms we were having last week, but that doesn't explain all their downtime.
______________________________________
And get their SIMS product, too, it's pretty well optimised for the high end.
_ ______________________________
According to what's currently coming down the grapevine, in the upcoming Sun/Netscape 'partnership' SIMS and Netscape's MTA products will be merged into one product, and sun will be replacing the buggy SIMS message store with the corresponding Messenger component. Hopefully this will make for one nice, solid server.
________________________________________
For 25000 users you want HIGH reliability. Any admin who put that many users on just one box has some headaches in store._ _____________________________
_________________________________________
I think there are enough horror stories about Exchange bouncing around that I don't need to comment on that. I recommend Netscape Messenger. It's designed to work with high-availability clustering solutions, has built-in IMAP capability, and is packaged with Netscape Directory Server which means all your user account data is in accessible through LDAP, and can be shared with other applications. You would definitely want to run it over Solaris w/Veritas or some other nice, reliable OS with a journaling filesystem.
_ ____________________________
Sun has its own messaging product, SIMS, which offers all of the same features. Don't do it. Half the bugs I have reported have been responded to sort of as follows: "Well, it just does that. I think they're going to release a patch for that as soon as they figure out why." I have not been impressed with the SIMS support team.
__________________________________________
But we have been having power problems, the jokers who wired our new server room screwed up big time. We've also lost 4 power modules in an APC Symmetra that was powering it. Can't fault Sun._ ___________________________
___________________________________________
According to crayz, we've all been scanning the same 115 work units over and over again. That's a grand total of about 30 megs of data. Seti's own computers should be able to handle that. What a pisser. Unfortunately I don't have a record of the work units my WinDoze boxes have run, but here are the ones my home Linux box has done. Notice that there are several duplicates. Do these look familiar to anyone else?
_ __________________________
Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency:
1.418886717 GHz
Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency:
1.420126951 GHz
Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency:
1.420126951 GHz
Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency:
1.420185545 GHz
Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency:
1.420439451 GHz
Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency:
1.420683592 GHz
Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency:
1.420869139 GHz
Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency:
1.420976560 GHz
Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency:
1.420976560 GHz
Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency:
1.420976560 GHz
Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency:
1.421142576 GHz
Sky coordinates: 10.5 R.A., 20.2 Dec Recorded on: Thu Jan 7 23:07:30 1999 Base Frequency:
1.421142576 GHz
Sky coordinates: 13.0 R.A., 12.7 Dec Recorded on: Fri Jan 8 00:51:05 1999 Base Frequency:
1.419101560 GHz
Sky coordinates: 13.0 R.A., 12.7 Dec Recorded on: Fri Jan 8 00:51:05 1999 Base Frequency:
1.419101560 GHz
Sky coordinates: 13.0 R.A., 12.8 Dec Recorded on: Fri Jan 8 00:49:44 1999 Base Frequency:
1.419404295 GHz
Sky coordinates: 13.1 R.A., 12.2 Dec Recorded on: Fri Jan 8 01:05:49 1999 Base Frequency:
1.419960936 GHz
____________________________________________
So what do we expect to find? Based on bandwidth usage on the Internet, my guess is that our first contact, once decoded, will turn out to be alien porno._ _____________________________
_________________________________________
On Linux whichever had the higher priority ('niceness') would win and probably mostly squash the other one. You could set the priority the same, then they would compete for CPU time, but why crunch keys when you could be looking for ET's TV signals?
_ _______________________________
Hmm, based on bandwidth usage on the Internet, I wonder what the probability is that our first contact will be picking up alien porno...
_______________________________________
Whoops, that should have been 78.750 GB per day._ ___________________________
___________________________________________
Last time I looked, there were over 525,000 people registered, and the average processing time for one 250,000 byte work unit is currently just under 40 hours (My P350 does one in about 20 hours). So, that works out to about .00625 MB per computer per hour. With 525,000 people registered, that's at least 3281.25 MB per hour being processed, or 78750 GB per day. More than twice as much data as they're pulling in from Arecibo. Of course, that's assuming all the computers are working on it 24 hours per day, but I'll bet most every computer is putting in at least 12 hours of CPU time on this per day. So, we're already processing the data faster than they are taking it in._ __________________________
____________________________________________
Because everyone is analyzing the same 30 megabytes of data. One Linux box I am running it on has processed 16 work units. FIVE of them have been duplicates.
_ __________________________________
It is a waste of bandwidth - with 525,000 users it's somewhere between 35 and 70 GB per day assuming their statistics are correct (average 40 hours per work unit), and that each of the 525,000 users put in between 12-24 hours of CPU time per day. I was putting in about 60 hours total CPU time per day until just now, and averaging about 20 hours per work unit, so these numbers are probably low.
I have stopped running the client on all my machines until I hear that they are sending out new data. Once I hear there's new data, I'll start up the clients again. I love the idea, I just wish they hadn't been so premature about releasing the clients and announcing the project everywhere.
____________________________________
Jane Dark's feature, "Suffer the (White, Middle-Class) Children" was a poorly researched, irresponsible piece obviously designed to be inflammatory. She basically calls the oppressed high-school geeks whiners. She says it is a white problem, and implies that these white folks don't know what suffering is.
_ __________________________
Well, she may be correct on the first count - in 1995 over 90% of suicides were by white folks. I think those people knew what suffering is. "Jane Dark" says she had her share of jerks to deal with in high school, and sure it sucked, but she was able to deal with it. I don't think she realizes the extent of the abuse some of these kids endure. Many of them are told daily that they are worthless human beings, a waste of oxygen. Maybe this doesn't happen as much in predominantly minority high schools. Whether or not that is true, the suffering exists, and it is extreme enough to drive these kids to kill themselves and each other. There is no point or purpose in making this a racial issue except to be inflammatory.
Jane, you have succeeded in provoking a response from me. I'm just sorry you had to belittle the suffering of thousands of teenagers to do it.
____________________________________________