Sample grand visions include how to create a 'self-sustaining, human presence throughout the solar system' and 'truly autonomous robotic operations for exploration and habitation.'
kudos! simply replacing part of the xhtml dom is exactly what is most interesting to many. Unfortunately, present online tutorials only show dropdown-menu exchanges, etc. as they are even simpler to implement.
what I personally would prefer is to use XSLt to translate machine-understandable XML (e.g., rdf) to html at the last stage of the pipeline. I'm pretty sure it's possible, but as webdev is not my forte I haven't looked into it yet.
It could seriously reduce bandwidth consumption at the cost of client-side CPU usage increase.
The No.1 library I'd like to see is some serverside code that automatically generates diffs between pages and the clientside to automatically apply the DOM changes. Diff support in HTTP was proposed (forgot the RFC no.), but never implemented. This could implement differential updates through a backdoor
It's the chop shops and small sellers that are cheating their customers...
which made me think. Why doesn't microsoft add a "where have you purchased this copy" question to their online registration process? That would probably scare the hell out of those chop shops. And rightly so, of course
sorry for the archive.org link, apparently tapster - like its lesser known brother - has succumbed to the power of the RIAA. Or was it because they streamed in dobly?
with more than 2100 posts already chances are noone is going to read this. Who cares, I'll just listen to myself then.
The presentation app, keynote, appears to be a godsend. It has a number of features that I always wanted (but was too lazy to code):
- a dual-monitor setup so that you can have a presentation on the beamer and an overview on your laptop. Do modern laptops carry dual-out, by the way?
- a timer to go with your overview page. The days that I have skipped content just to fit the deadline are nearly over. Finally.
however, there is one feature that I'm still missing. This one is especially useful for technical design, etc: construct individual slides from `master' images that are possibly larger than the slides. In a CAD environment it means flying into a detail of your design.
In general this technique should lead to a more natural progression from slide to slide. Perhaps it can be generalized even. I'm thinking along the lines of first creating a story and only afterwards chopping it up into bytesize chunks. The aforementioned design-issue is just an example. Read "presenting to win" by Weissman (yeah, horrible title) for more useful comments on holding presentations.
from the original blurb: automatically tune themselves for the best possible performance for any given workload
as I understand it, and as you mention, GAs try to converge a solution to a global optimum. However, it appears that they're trying to use it in a situation where the optimal solution is a moving target.
While the idea is great I'm pretty sure that there must be more applicable optimization strategies than GAs for this problem. While on the topic of AI techniques I would suggest simple neural nets (feed-forward with back-propagation), as these tend to quickly adapt to changes in variables such as workload.
NB: my knowledge on the topic is limited to a single course a few years back.
I could be wrong here, but isn't gentoo the only distribution that allows you to experiment with packets in such a way that you can easily remove them and their dependencies?
emerge unmerge [ANYTHING] is great. Coupled with an emerge depclean for removal of packages that were installed purely for satisfying dependencies it greatly reduces the dependency hell.
With rpm/apt you can undoubtedly remove a single package, but what about the depclean feature? is it there?
in any case, this was one of my main reasons for switching to gentoo two years ago. No more need for clean installs because I forgot where all the junk came from.
Oh, another reason was that I had to keep my 200Mhz machine running through my thesis work; the performance improvement could be a placebo effect, but it worked when I needed it:)
network processors are basically what you're asking for. And guess what? Intel has a platform called IXP (based on the IXA design). Many other companies supply network processor platforms as well, by the way.
Right now, NPs are basically research material, but I think that you can expect to see some of these NIC+logic boards surface in mainstream computing devices in the near future.
The reason that I know this is that I have been part of a research initiative in packet filtering on IXPs, albeit for monitoring purposes. Ref: ffpf.sf.net
Is it normal for companies like SGI and Cray to have `oil and gas industry analysists' or a 'Director of Complex Scientific Computing Projects'? The only site referencing the latter term is his, not SGI's.
I'm not saying the CV is a fake, but some of it I just don't understand. For someone that has worked at Cray's Research departement and was the aforementioned director at SGI he hasn't published a lot. Citeseer doesn't know him, neither does DBLP. To be fair, publications in french wouldn't be indexed. Then again, he (allegedly) worked for american companies.
Google has a few pages with his emailadres @cray from '95, but that doesn't say the address is correct. AFAIK Cray has never had a french research departement.
So, is this harmless resume embellishment or more? Hopefully I'm just getting paranoid. Roland, if you're listening, perhaps you can explain (or update your CV).
What I hate is Sourceforge's prdownload stuff that has you getting through all that then doing a redirect to force a browser-based download. I wish they wouldn't do that.
there's a way around this problem. I once wrote an auto-update feature for my sourceforge hosted project that relied on wget.
the trick is to preselect their local server, OSDL.dl:
http://osdn.dl.sourceforge.net/[your files]
visit ffpf.sf.net and click on "download latest" for a proof of concept.
By the way, I posted this hack to the sourceforge feature request list, but they never got back to me.
I recently read a PhD dissertation discussing (among others) a java-based grid environment. At first the entire idea sounded odd to me because of an expected performance drop. However, cases where java can be as fast as and sometimes even faster than optimized C have been shown before.
In the dissertation, two papers are mentioned that compare java to C and apparently show positive results for java. I haven't actually read these papers, by the way. However, their titles suggest that they might be on-topic here.
As citeseer is always crawling, even without a link on slashdot, I was afraid to post this... but then I remembered that noone here actually reads the FM:) Jokes aside, it is a great resource and I hope that they get some mirrors soon.
as there are already some 400 posts, perhaps noone's going to read this. Still, here's my two cents.
Randall states that CS students don't learn assembly language anymore, or that if they do, they aren't being taught well. I have to disagree.
In our CS curriculum (Universiteit Leiden, NL) we had courses that learned us
(1) how to design a CPU, including ALU, load/store architecture and microinstructions for a fake CPU that interprets (a subset of) the MIPS instruction set. (digital systems design)
(2) write a compiler from a subset of pascal to MIPS code. The end result had to be tested on SGI Indy's. Present students still have to do this, but need SPIM, since the Indy's have gone, how sad. (compiler construction *)
My point is that our curriculum is arguably representative of the standard CS bachelor, at least in Europe, and that learning everything from high-level languages up to micro-instructions greatly surpasses dabbling with programming in assembly.
I do agree with his main point though: you should know these things. My feeling is that more people are already acquinted with Randall's dark art than he thinks.
(*) offtopic: this course was the major stumbling point of half of all CS students. Some handed in the assignments 4 years late, after their final MSc paper.
It appears that either you are good at the math side and suck at coding or it is the other way around. I did fairly well here, but I won't go into my math;)
Although there are already some 300 posts I haven't read one success story yet (then again, I read at +5).
I'm actually extremely happy with the current situation, and use the two mechanisms at the same time. For instance, when I'm editing a text or sourcefile, I use ctrl-x/ctrl-v for moving around large portions of the code. As long as I havened pasted, the original order is still in tact.
Then, using the X-windows selection mechanism, I first clean up the new location before pasting. Especially in large files this saves a lot of scrolling through the text.
nb: perhaps others have written similar stories, but then they weren't modded up till 5: interesting. I only read about 30 `I can't use it so it must be bad stories'. What does this say about the/. crowds? Would the real geeks please stand up;)
Oh, and ofcourse you can always choose not to use either of the mechanisms, so please stop complaining. The fact that you have the choice is good. I hate posts saying that this is why F/OSS / Linux / insert keyword here hasn't broken through on the desktop yet.
While the interview is light on details, there is more information available online.
Don't forget how the system works. Darpa basically hands out money for research into areas it finds interesting. Coincidently, I've been involved for a short time in a research project dealing with exchanging present day IP (mostly the heavyweight gorilla listening to the name TCP) with smaller, more adaptable alternatives.
The basic idea, AFAIK, is to do away with the one size fits all model of networking and replace it with a more adaptive lego-like stack. For this to work you need information on the state of the network in order to build your optimal dynamic stack. A possible source for this might be the discussed knowledge plane. Also, actual micro-protocols need to be created and some sort of decision making system must be in place (APnets). Shameless plug of my own work here.
I don't know of other projects, but if Darpa has opened its wallet for this cause you can expect many other universities to have similar initiatives underway.
apart from these there are many more interesting links, but mostly those have to do with specialty domains, such as networking or memory management. I pasted my bookmarks.html to my personal website.
And lastly, if you want to do cross-version/dev (mknod and devfs), mmap, module or networking development, please take a look at our project's cvs server through our website (webcvs) at ffpf.sourceforge.net. The directory ffpf/srv/v1.1/generic contains some files I created that work on both 2.4.24 and 2.6.1 and which I'll test on a 2.3.99 system shortly.
luckily they stopped him... now london was only devastated in the big fire of 1666.
In retrospect, it didn't really matter. If London would have been blown up in '05, the newly built brick houses wouldn't have been swept by the flames a few decades later.
Slashdot Mirror
Sample grand visions include how to create a 'self-sustaining, human presence throughout the solar system' and 'truly autonomous robotic operations for exploration and habitation.'
for 50 to 75K? That's K as in years, I hope.
kudos! simply replacing part of the xhtml dom is exactly what is most interesting to many.
Unfortunately, present online tutorials only show dropdown-menu exchanges, etc. as they are even simpler to implement.
what I personally would prefer is to use XSLt to translate machine-understandable XML (e.g., rdf) to html at the last stage of the pipeline. I'm pretty sure it's possible, but as webdev is not my forte I haven't looked into it yet.
It could seriously reduce bandwidth consumption at the cost of client-side CPU usage increase.
The No.1 library I'd like to see is some serverside code that automatically generates diffs between pages and the clientside to automatically apply the DOM changes. Diff support in HTTP was proposed (forgot the RFC no.), but never implemented. This could implement differential updates through a backdoor
moi, pour un, acceuill nos nouveaux overlords Redmond.
(yeah, the french stinks. I'm taking lessons)
if this doesn't shut up all those pragmatists shouting that transparency is just useless eye-candy, then I don't know what will.
Now I can finally stare at the wall all day long.
why not just read the english version of al jazeera?
Actually, Java would be perfect for managing a grid.
:)
If you think so, check it ibis. You can download the latest version from the link and play around with it. Or read this old slashdot story
disclaimer: I'm not directly involved with the project, but working in the same group as the developers. And I don't mind pushing a good idea
It's the chop shops and small sellers that are cheating their customers ...
which made me think. Why doesn't microsoft add a "where have you purchased this copy" question to their online registration process? That would probably scare the hell out of those chop shops. And rightly so, of course
They are aiming particularly to tap the younger video-game generation
.
I think they misspelled tapster.com
Background story>
sorry for the archive.org link, apparently tapster - like its lesser known brother - has succumbed to the power of the RIAA. Or was it because they streamed in dobly?
with more than 2100 posts already chances are noone is going to read this. Who cares, I'll just listen to myself then.
The presentation app, keynote, appears to be a godsend. It has a number of features that I always wanted (but was too lazy to code):
- a dual-monitor setup so that you can have a presentation on the beamer and an overview on your laptop. Do modern laptops carry dual-out, by the way?
- a timer to go with your overview page. The days that I have skipped content just to fit the deadline are nearly over. Finally.
however, there is one feature that I'm still missing. This one is especially useful for technical design, etc: construct individual slides from `master' images that are possibly larger than the slides. In a CAD environment it means flying into a detail of your design.
In general this technique should lead to a more natural progression from slide to slide. Perhaps it can be generalized even. I'm thinking along the lines of first creating a story and only afterwards chopping it up into bytesize chunks. The aforementioned design-issue is just an example. Read "presenting to win" by Weissman (yeah, horrible title) for more useful comments on holding presentations.
from the original blurb:
automatically tune themselves for the best possible performance for any given workload
as I understand it, and as you mention, GAs try to converge a solution to a global optimum. However, it appears that they're trying to use it in a situation where the optimal solution is a moving target.
While the idea is great I'm pretty sure that there must be more applicable optimization strategies than GAs for this problem. While on the topic of AI techniques I would suggest simple neural nets (feed-forward with back-propagation), as these tend to quickly adapt to changes in variables such as workload.
NB: my knowledge on the topic is limited to a single course a few years back.
Quite an eloquent piece. You were properly countering the stereotype (though too elitist IMHO). Until I read your nickname...
Beardo, my man, I guess you're the guy that started all the stories and you don't even know it.
I could be wrong here, but isn't gentoo the only distribution that allows you to experiment with packets in such a way that you can easily remove them and their dependencies?
:)
emerge unmerge [ANYTHING] is great. Coupled with an emerge depclean for removal of packages that were installed purely for satisfying dependencies it greatly reduces the dependency hell.
With rpm/apt you can undoubtedly remove a single package, but what about the depclean feature? is it there?
in any case, this was one of my main reasons for switching to gentoo two years ago. No more need for clean installs because I forgot where all the junk came from.
Oh, another reason was that I had to keep my 200Mhz machine running through my thesis work; the performance improvement could be a placebo effect, but it worked when I needed it
pornographic images stay in the brain forever.
dare to test this hypothesis?
network processors are basically what you're asking for. And guess what? Intel has a platform called IXP (based on the IXA design). Many other companies supply network processor platforms as well, by the way.
Right now, NPs are basically research material, but I think that you can expect to see some of these NIC+logic boards surface in mainstream computing devices in the near future.
The reason that I know this is that I have been part of a research initiative in packet filtering on IXPs, albeit for monitoring purposes. Ref: ffpf.sf.net
Well, his resume is NOT extraordinary.
actually, his
CV does look quite amazing.
Is it normal for companies like SGI and Cray to have `oil and gas industry analysists' or a
'Director of Complex Scientific Computing Projects'? The only site referencing the latter term is his, not SGI's.
I'm not saying the CV is a fake, but some of it I just don't understand. For someone that has worked at Cray's Research departement and was the aforementioned director at SGI he hasn't published a lot. Citeseer doesn't know him, neither does DBLP. To be fair, publications in french wouldn't be indexed. Then again, he (allegedly) worked for american companies.
Google has a few pages with his emailadres @cray from '95, but that doesn't say the address is correct. AFAIK Cray has never had a french research departement.
So, is this harmless resume embellishment or more? Hopefully I'm just getting paranoid. Roland, if you're listening, perhaps you can explain (or update your CV).
What I hate is Sourceforge's prdownload stuff that has you getting through all that then doing a redirect to force a browser-based download. I wish they wouldn't do that.
there's a way around this problem. I once wrote an auto-update feature for my sourceforge hosted project that relied on wget.
the trick is to preselect their local server, OSDL.dl:
http://osdn.dl.sourceforge.net/[your files]
visit ffpf.sf.net and click on "download latest" for a proof of concept.
By the way, I posted this hack to the sourceforge feature request list, but they never got back to me.
I recently read a PhD dissertation discussing (among others) a java-based grid environment. At first the entire idea sounded odd to me because of an expected performance drop. However, cases where java can be as fast as and sometimes even faster than optimized C have been shown before.
In the dissertation, two papers are mentioned that compare java to C and apparently show positive results for java. I haven't actually read these papers, by the way. However, their titles suggest that they might be on-topic here.
The first is "Benchmarking Java against C and Fortran for Scientific Applications" by J. M. Bull, L. A. Smith, L. Pottage and R. Freeman
Second, and judging from its title perhaps less on-topic is "High-Performance Java Codes for Computational Fluid Dynamics" by Christopher Riley, Siddhartha Chatterjee and Rupak Biswas
finally, the dissertation can be found on Rob van Nieuwpoort's publication list and is titled "Efficient Java-Centric Grid-Computing"
As citeseer is always crawling, even without a link on slashdot, I was afraid to post this... but then I remembered that noone here actually reads the FM :) Jokes aside, it is a great resource and I hope that they get some mirrors soon.
this article was the first I read for a long time to have the dot com boom's `zingy' feel to it again.
I guess it's time to start going back to reading Red Herring and buying stocks in useless companies...
I, for one, can't wait.
as there are already some 400 posts, perhaps noone's going to read this. Still, here's my two cents.
;)
Randall states that CS students don't learn assembly language anymore, or that if they do, they aren't being taught well. I have to disagree.
In our CS curriculum (Universiteit Leiden, NL) we had courses that learned us
(1) how to design a CPU, including ALU, load/store architecture and microinstructions for a fake CPU that interprets (a subset of) the MIPS instruction set. (digital systems design)
(2) write a compiler from a subset of pascal to MIPS code. The end result had to be tested on SGI Indy's. Present students still have to do this, but need SPIM, since the Indy's have gone, how sad.
(compiler construction *)
My point is that our curriculum is arguably representative of the standard CS bachelor, at least in Europe, and that learning everything from high-level languages up to micro-instructions greatly surpasses dabbling with programming in assembly.
I do agree with his main point though: you should know these things. My feeling is that more people are already acquinted with Randall's dark art than he thinks.
(*) offtopic: this course was the major stumbling point of half of all CS students. Some handed in the assignments 4 years late, after their final MSc paper.
It appears that either you are good at the math side and suck at coding or it is the other way around. I did fairly well here, but I won't go into my math
it's spelled `philips', not phillip's or phillips. Just look at the URL.
Although there are already some 300 posts I haven't read one success story yet (then again, I read at +5).
/. crowds? Would the real geeks please stand up ;)
I'm actually extremely happy with the current situation, and use the two mechanisms at the same time. For instance, when I'm editing a text or sourcefile, I use ctrl-x/ctrl-v for moving around large portions of the code. As long as I havened pasted, the original order is still in tact.
Then, using the X-windows selection mechanism, I first clean up the new location before pasting. Especially in large files this saves a lot of scrolling through the text.
nb: perhaps others have written similar stories, but then they weren't modded up till 5: interesting. I only read about 30 `I can't use it so it must be bad stories'. What does this say about the
Oh, and ofcourse you can always choose not to use either of the mechanisms, so please stop complaining. The fact that you have the choice is good. I hate posts saying that this is why F/OSS / Linux / insert keyword here hasn't broken through on the desktop yet.
http://ffpf.sourceforge.net/
includes libpcap support, but is more extensible (filter language bindings) and efficient (kernelspace & userspace processing).
btw: I'm one of the developers.
While the interview is light on details, there is more information available online.
Don't forget how the system works. Darpa basically hands out money for research into areas it finds interesting. Coincidently, I've been involved for a short time in a research project dealing with exchanging present day IP (mostly the heavyweight gorilla listening to the name TCP) with smaller, more adaptable alternatives.
Two projects in this field that I've heard of
are
the knowledge plane and
application private networks
The basic idea, AFAIK, is to do away with the one size fits all model of networking and replace it with a more adaptive lego-like stack. For this to work you need information on the state of the network in order to build your optimal dynamic stack. A possible source for this might be the discussed knowledge plane. Also, actual micro-protocols need to be created and some sort of decision making system must be in place (APnets). Shameless plug of my own work
here.
I don't know of other projects, but if Darpa has opened its wallet for this cause you can expect many other universities to have similar initiatives underway.
I couldn't agree more. A month ago I started serious kernel programming and my code wouldn't have been what it is if I hadn't found out about the free, online version of the second edition.
:
/dev (mknod and devfs), mmap, module or networking development, please take a look at our project's cvs server through our website (webcvs) at ffpf.sourceforge.net. The directory ffpf/srv/v1.1/generic contains some files I created that work on both 2.4.24 and 2.6.1 and which I'll test on a 2.3.99 system shortly.
Some other useful sites
the Linux kernel API reference
Linux cross reference, especially the `identifier search'.
the driver porting series over at LWN.net (which Rufus211 alreay pointed out).
the module init tools FAQ
this document, aptly titled `the linux kernel'
apart from these there are many more interesting links, but mostly those have to do with specialty domains, such as networking or memory management. I pasted my bookmarks.html to my personal website.
And lastly, if you want to do cross-version
luckily they stopped him... now london was only devastated in the big fire of 1666.
In retrospect, it didn't really matter. If London would have been blown up in '05, the newly built brick houses wouldn't have been swept by the flames a few decades later.