Gee, another hardware company weighs in behind Free Software. The folks who crank out millions of blinkey-light boxes see hard economics at work here: Use Linux, don't have to hire expensive code-boys to worry about which O(1) scheduler to use. Can even afford to pitch in to the effort through places like OSDL with the MILLIONS saved in software development costs.
SATPO*: Solaris coders are a dying breed, costing their mother company big bucks to produce something that can be had for free, and they can't even think about dragging their company to the Linux bandwagon until SCO self-immolates. Solaris will take big hits when 1) some Linux or *BSD gets a DII-COE certificate (DoD apps), and 2) CGL turns 1.0.
I think the NTT support of OSDL is great, but I'm not going to pat them on the back for being philantropic - they did this for cold, hard economic reasons.
When I do projects, I almost always end up with a combination of parts from three sources:
CPUs from some special vendor;
Radio Shack for those goofy little PC boards and miscellaneous hardware;
OEM Parts for everything else.
Mo' better organized now than in the early '90s, and Dick is still hand-writing invoices. Recently I had a hard time finding 40-pin DIP sockets, 'till I dropped in on the folks on Hancock St north of Fillmore...
What needs to happen in this discussion is a separation of the need to manage a college's IT infrastructure vs. the flexibility to teach sound computer science. The two are (usually) mutually exclusive.
I used to chair a computer science department for a small private university. I had a M$-equipped box on my desk, but almost all of my teaching involved Linux or Solaris computers. Others taught with the appropriate tools to fill the need; MCSC courses naturally used M$ stuff. But we had the freedom to use whatever tools met the need, which I think makes for a more diverse and enriching experience for students. At the same time, we recognized the need for the IT guys to manage administrative systems based on a consistent solution, and for them it was M$.
For all you CS students out there complaining that you're not getting enough M$ or whatever, get over it! You're much more valuable to me as a prospective employer with a solid foundation in the computing sciences, sans vendor specifics: with that, you can buy the O'Reilly book to learn the latest and greatest, rather than taking more classes (boy, the admissions advisors hated for me to say stuff like that!).
Be it OSS or M$, what would be sad is for a school to bind their _academic_ computing infrastructure to either one. But the IT guys need to manage efficiently, or your tuition will go up (oh, it is anyway.....?)
Okay, do you think it's a good idea to shoehorn al this media stuff into the IP(V6) layer, or does an Internet-scale ATM service make more sense? (Ooh, I guess I've telegrahped my druthers here...:-) Anyway, what are your inclinations toward an architecture for information services?
>... tried to duplicate the familiar WindowsXP UI > feel under Linux.
Familiar? Hmph. I've been dorking with my wife's new HP laptop running XP for the past week, trying to get TCP/IP bound to the wireless LAN card. Took a usenet search to reveal that the bindings are gotten under a_menu pick and not in the pop-up for the interface (Lordeee, save us all!!!). Why do they feel that they have to keep moving things around, except to justify their self-perpetuating training program?
Soooo, I'm copying crap off my Win98 laptop to my wife's XP laptop so I can burn CDs of it and make room to install a Linux which I'll then boot and run Win98 when needed under VMWare. Not a solution for everyone, but I feel like I have more control in a Unix OS.
In one of my doctorate classes, I did an informal study of the relationship between software size and the accuracy of its COCOMO estimate on a data set of government projects. In short, I found that the model was _completely_ wrong estimating the duration of projects under 10,000 lines of code, +/- 20% for projects between 10,000 and about 200,000 lines of code, and it got really accurate after that. Just went looking for the paper but couldn't find it, so the above assertion is only a recollection. Point is, it's probably easier to be "close enough" on really big projects, and don't waste your time on small ones. I might try to get one of my grad students to duplicate or redo the study with a better project set.
Re:While On The Subject Of Laptops & Linux...
on
Which Laptop To Buy?
·
· Score: 1
The compatibility among products is quite high; most any will probably work. I'd check with the school first, see if they have a list of compatible cards. Otherwise, the Linksys and D-link products are cheap and functional. I currently have a Thinkpad 240X with a Aircomnet card for use at work; works great.
I use VNC to admin W2K servers both from on-site and from home. At home, it's 640K DSL over about 9 hops; it's a little sluggish, but I've spent hours doing server configs without running from the room screaming. I've even rebooted from home; comes back up just fine. I've had to turn on "entire-screen polling" in order to get some MS apps to update the screen, but even that doesn't kill the experience. Didn't know about TightVNC until this thread; I'm going to try it tonight!
I did it once from my in-laws over a modem to show my father-in-law what kind of work we're doing - he was more impressed with VNC than my day job. Oh well...
Forget ./, What Happens to Internet during War?
on
Slashdot During War?
·
· Score: 1
A while ago, ran across a group of folks who ping the entire internet on a regular schedule, then produce maps of their results. They have a page showing the adaptation that took place during a part of the Serbian conflict:
Now, I don't know what'd happen to./, but if there's connectivity, someone will step in. I've got my old ham rig boxed in the basement, just in case; that, and a packet radio TNC, and I could be a post-armegeddon ISP!
They did publish an abbreviated version of the protocol that you can get at their site in.pdf form. http://www.digitrax.com/loco1hdr.htm
I've tried to program to it; if you're used to layered protocols, it's not trivial. I participated in a discussion to build a Linux device driver for it, but guys who know a lot more about DDs than I got frustrated...
I think that Digitrax is going to see the wisdom of opening the entire protocol before long, so that conformant products take off and stimulate their business in turn.
Slashdot Mirror
Gee, another hardware company weighs in behind Free Software. The folks who crank out millions of blinkey-light boxes see hard economics at work here: Use Linux, don't have to hire expensive code-boys to worry about which O(1) scheduler to use. Can even afford to pitch in to the effort through places like OSDL with the MILLIONS saved in software development costs.
SATPO*: Solaris coders are a dying breed, costing their mother company big bucks to produce something that can be had for free, and they can't even think about dragging their company to the Linux bandwagon until SCO self-immolates. Solaris will take big hits when 1) some Linux or *BSD gets a DII-COE certificate (DoD apps), and 2) CGL turns 1.0.
I think the NTT support of OSDL is great, but I'm not going to pat them on the back for being philantropic - they did this for cold, hard economic reasons.
*SATPO - Someone Arbitrary To Pick On
When I do projects, I almost always end up with a combination of parts from three sources:
Mo' better organized now than in the early '90s, and Dick is still hand-writing invoices. Recently I had a hard time finding 40-pin DIP sockets, 'till I dropped in on the folks on Hancock St north of Fillmore...
I can't remember the name, but it was on Mississippi...
What needs to happen in this discussion is a separation of the need to manage a college's IT infrastructure vs. the flexibility to teach sound computer science. The two are (usually) mutually exclusive.
I used to chair a computer science department for a small private university. I had a M$-equipped box on my desk, but almost all of my teaching involved Linux or Solaris computers. Others taught with the appropriate tools to fill the need; MCSC courses naturally used M$ stuff. But we had the freedom to use whatever tools met the need, which I think makes for a more diverse and enriching experience for students. At the same time, we recognized the need for the IT guys to manage administrative systems based on a consistent solution, and for them it was M$.
For all you CS students out there complaining that you're not getting enough M$ or whatever, get over it! You're much more valuable to me as a prospective employer with a solid foundation in the computing sciences, sans vendor specifics: with that, you can buy the O'Reilly book to learn the latest and greatest, rather than taking more classes (boy, the admissions advisors hated for me to say stuff like that!).
Be it OSS or M$, what would be sad is for a school to bind their _academic_ computing infrastructure to either one. But the IT guys need to manage efficiently, or your tuition will go up (oh, it is anyway.....?)
Okay, do you think it's a good idea to shoehorn al this media stuff into the IP(V6) layer, or does an Internet-scale ATM service make more sense? (Ooh, I guess I've telegrahped my druthers here... :-) Anyway, what are your inclinations toward an architecture for information services?
> ... tried to duplicate the familiar WindowsXP UI
> feel under Linux.
Familiar? Hmph. I've been dorking with my wife's new HP laptop running XP for the past week, trying to get TCP/IP bound to the wireless LAN card. Took a usenet search to reveal that the bindings are gotten under a_menu pick and not in the pop-up for the interface (Lordeee, save us all!!!). Why do they feel that they have to keep moving things around, except to justify their self-perpetuating training program?
Soooo, I'm copying crap off my Win98 laptop to my wife's XP laptop so I can burn CDs of it and make room to install a Linux which I'll then boot and run Win98 when needed under VMWare. Not a solution for everyone, but I feel like I have more control in a Unix OS.
In one of my doctorate classes, I did an informal study of the relationship between software size and the accuracy of its COCOMO estimate on a data set of government projects. In short, I found that the model was _completely_ wrong estimating the duration of projects under 10,000 lines of code, +/- 20% for projects between 10,000 and about 200,000 lines of code, and it got really accurate after that. Just went looking for the paper but couldn't find it, so the above assertion is only a recollection. Point is, it's probably easier to be "close enough" on really big projects, and don't waste your time on small ones. I might try to get one of my grad students to duplicate or redo the study with a better project set.
The compatibility among products is quite high; most any will probably work. I'd check with the school first, see if they have a list of compatible cards. Otherwise, the Linksys and D-link products are cheap and functional. I currently have a Thinkpad 240X with a Aircomnet card for use at work; works great.
I did it once from my in-laws over a modem to show my father-in-law what kind of work we're doing - he was more impressed with VNC than my day job. Oh well...
http://plan9.bell-labs.com/who/ches/map/yu/index.h tml
Now, I don't know what'd happen to ./, but if there's connectivity, someone will step in. I've got my old ham rig boxed in the basement, just in case; that, and a packet radio TNC, and I could be a post-armegeddon ISP!
I've tried to program to it; if you're used to layered protocols, it's not trivial. I participated in a discussion to build a Linux device driver for it, but guys who know a lot more about DDs than I got frustrated...
I think that Digitrax is going to see the wisdom of opening the entire protocol before long, so that conformant products take off and stimulate their business in turn.