Slashdot Mirror

Although IBM http://www.research.ibm.com/ may be out of the disk drive business, they are still working on it. Take a look at the Almaden Research Center in San Jose http://www.almaden.ibm.com/ still going strong after all these years.

Now that you mentioned SCIAM.
There is an article in the august issue of Scientific American about magnetologic gates. This mentions that instead of making transistors smaller so you can put more of them in the same space. You could also try achieve the same functions using less elements.

magnetologic gates are based on the MRAM technology. With some modifications the designs for MRAM can be used to create logic gates that are much more efficient and powerfull then CMOS based transistors.
With only 1 magnetologic gate you could create a AND, OR, NOR or NAND function. with 2 gates you can create a XOR function with would require 8 to 14 CMOS transistors. The 'full adder', the most used unit in a processor used to add two binary inputs, can be created with only 3 gates instead of 16 CMOS transistors.
So using magnetologic gates you can achieve the same kind of processing power improvement without using smaller units.

These magnetologic gates have some other advantages. They are non-volatile so they remember/store the result of the last calculation performed and reading out this value does not delete the information. This means that the overall calculation can be performed faster and it also enables parallel or clockless execution of operations.

Magnetologic gates can be reprogrammed like FPGA's. But unlike FPGA's switching between different functionalities takes just billions of a second. This ability to morph (which is the main focus of this article) radically reduces the amount of transistors needed in a processor. Since all function are hardwired in a normal CMOS processor, at any given time only a few percent of the transistors are actually used. If you could change the function of your elements with every operation, you could perform the same scala of different funtions with just a few elements.

If this technology will progress it could bypass the miniaturization efforts.

My college Senior Project was a Linux-based Domino server implementation. Worked great for me.

Click here to download a trial of Lotus Domino for Linux. Click here for a Notes client for Windows (works on Wine) and Mac.

My college Senior Project was a Linux-based Domino server implementation. Worked great for me.

Click here to download a trial of Lotus Domino for Linux. Click here for a Notes client for Windows (works on Wine) and Mac.

IBM was the contractor, so you can't really fault them for SOCOG not paying for accessibility. I dobut any company would just throw it in for free! I spent a year at IBM working mostly on accessibility issues with internal and external sites, it's something they've paid a lot of attention to.

It's simple enough to add alt text to all images, but ensuring things are fully accessible is a lot more work. CSS makes this a lot easier than with table layouts though.

Why buy a 360 when you can buy a 720
?

Firefox already adheres to standards better than IE, has a more rubust, and secure environment, and arguably provides a superior user experience to IE, and yet IE lives on... So why would some (arguably nice) DHTML addons make a difference?

Many organisations are legally obliged to make their internal applications, including web applications, accessible to the disabled.

Two of the most popular applications that can read websites out to blind people, JAWS and IBM Homepage Reader, are both based on Internet Explorer.

This code will supposedly make it easier for web applications using DHTML to work in Firefox for disabled people.

Usually IBM has got good code

It's worth noting that IBM built one of the most popular browsers for blind people, IBM Homepage Reader, which is currently based on Internet Explorer. Perhaps this is a move to help them switch to Firefox in the future?

I can't help but point out though, that in a landmark website accessibility case, SOCOG were fined A$20,000 for not having an accessible website for the Sydney Olympics. Guess who built their website? Yup, it was IBM :).

The issue of Software patents is a touchy one, and MS is desparate to have it approved. They even went as far as blackmailing the government of Denmark. They know that it wouldn't be constructive to give any extra ammo to it's opponents at such a critical time.

The goal of MS is to subsume OpenSource or extinguish it. Remember the failed MS email standard that contained both a submarine patent and licensing that strictly forbid GPL developement? MS allowed that technology to die stillborn rather than bend and allow GPL use of any of it's patents (as IBM does, see here).

Once software patent laws are in place and enforcible, do you honestly think that MS would not use Software Patents to toast the one competitor it could never control/buy/extinguish, Open Source and the GPL?

It's funny that this topic should come up on Slashdot today. Just a few hours ago, I was wrapping up my IBM Exterme Blue internship by a final session at Corporate Headquarters in Armonk, New York where teams from accross North America presented their projects to executives. We got to talk to, and dine with, some very interesting people from very high up in the company, both from the technical and business sides of the company. A detailed description of the internship is on http://ibm.com/extremeblue. Basically, it's a very fast-paced internship where teams of 3 technical students and 1 MBA student put together a new project from a very vague topic provided to them at the beginning of the 14-week term. The mentors who come up with the ideas and guide the projects are very influential and experienced people within the company, so the projects are truly important and innovative instead of being just "toys". Because each project is like a mini-product, we got to see all parts of IBM, not just the different technical departments we were working with but also the business side and even clients, and we managed to accomplish real work with each of them. The actual projects are very strongly aligned with IBM's goals, but they vary quite a bit since the company is large; they include new features for existing products, new solutions, research, client consulting, and even open-source (several involved Linux or Eclipse). The internship was very hectic, but it was an excellent way to see how a project goes through an entire development cycle at a big company, how the big decisions are made as it goes along, and most interesting to me, as a tech intern, how the business side of the company works.

Sera

Sorry, but I have to correct you on this one:
"Cheaper" is not IBM's key argument; in fact, every presentation about IBM's software strategy stresses the fact that open standards and open source lie at the core of the strategy to eliminate dependencies and create freedom to act.
Please read corporate strategies and look at the ongoing activities before posting a rant about them.
Take a look at http://www-306.ibm.com/software/info/topic/ if you want to learn more, for instance in articles like "Beyond TCO - The Unanticipated Second Stage Benefits of Linux" focuses on business applications and mission critical workloads on Linux and the often unanticipated second stage benefits that leading edge Linux users enjoy today.

As far as I can tell, IBM still sells workstations. We have about 20 ourselves (albeit, they are ~3 years old). Each of them (but one) have new power supplies. They all failed within the first 24 months. Sigh...

Wrong: "Cheaper" is not IBM's key argument; in fact, every presentation about IBM's software strategy stresses the fact that open standards and open source lie at the core of the strategy to eliminate dependencies and create freedom to act.
Please read corporate strategies and look at the ongoing activities before posting a rant about them.
Take a look at http://www-306.ibm.com/software/info/topic/ if you want to learn more, for instance in articles like "Beyond TCO - The Unanticipated Second Stage Benefits of Linux" focuses on business applications and mission critical workloads on Linux and the often unanticipated second stage benefits that leading edge Linux users enjoy today.

With a ThinkPad, it's pretty easy. Just get the UltraBay HDD caddy, and put the HDD in. Or, if you've got a desktop lying around, 40 pin to 44 pin IDE convertors are cheap.

Parts do cost a lot more, unless you buy from a third party. The advanced laptop changes are diagrammed EXTREMELY well with exploded diagrams and good notes along the way on ThinkPads - just take a look at this: ftp://ftp.software.ibm.com/pc/pccbbs/mobiles_pdf/2 4p2975.pdf
Scroll down to the section on FRU replacement.

As for permanent DVR solutions, that's where Mini-ITX (well, Mini-ITX Express boards, which are made by Commell) shines. I wouldn't use a lappy as a permanent DVR. A mobile one MAYBE, but not permanent.

Also a lot of this info is probably contained in IBM's Red Books: http://www.redbooks.ibm.com/cgi-bin/searchsite.cgi ?query=linux+performance

I'll admit it: I am an intern. As a matter of fact, I'm one of the best interns money can buy. (oooooh look shiny resume). But it's only a summer internship and has nothing repeat NOTHING (sort of quasiobligatory disclaimer here in case someone IBM is watching) to do with my work with this web site management position, which I mostly do during the school year (and is officially some sort of part-time salaried position with a university's economics department, if I recall correctly).

All you need to know.

I purchased a T21 just before going to college last year. At the time, it cost me just over $600. Now I could get the same refurbished Thinkpad for under $500.

It's a 800 Mhz machine but it is built like a tank. I just wanted a laptop for taking notes and working away from my dorm room. I don't need a fast laptop because I do most of my work on my Athlon 64 desktop system.

You still have to buy the software that will plug into the framework in order to actually process the information, though some open source projects are certain to come along.

This is interesting stuff, but not as thrilling as the article would suggest. Imagine if google open sourced their systems software, except the part that does the whole PageRank thing.

After searching for a whooping 5 minutes and even googling (gasp!) I couldn't find any decent article about what this actually is, just lots of info on how to use it. It looks like there is a new query language so it might be interesting for query expansion. But how does it extract these key facts from the documents? Does it do real natural language analysis? Just guess by looking at the document terms like every other search technology? Or is it just a framework that doesn't really do anything by itself? It sure looks like it when skimming http://www.research.ibm.com/journal/sj/433/gotz.ht ml so no revolution yet, sorry.

Definitely read/skim the SDK User's Guide http://dl.alphaworks.ibm.com/technologies/uima/UIM A_SDK_Users_Guide_Reference.pdf

The annotator premise is almost too simple; it's brilliant.

Looks interesting, there's a blurb about it here:
http://www-306.ibm.com/software/data/umm/umm.html

You can dowload the Unstructured Information Management Architecture SDK from alphaworks and take a good look at how to analyze unstructured information (text, audio, video, images, etc.) to discover, organize, and deliver relevant knowledge.

It's available now. As the article says:

UIMA technology is expected to be made available through open-source software site SourceForge by the end of 2005. The UIMA framework can currently be downloaded free of charge from IBM AlphaWorks at http://www.alphaworks.ibm.com/tech/uima/.

So, I ask, why wait for it to appear on SF if we can get it now?

It's available now. As the article says:

UIMA technology is expected to be made available through open-source software site SourceForge by the end of 2005. The UIMA framework can currently be downloaded free of charge from IBM AlphaWorks at http://www.alphaworks.ibm.com/tech/uima/.

So, I ask, why wait for it to appear on SF if we can get it now?

The important information is simply the url http://www.alphaworks.ibm.com/tech/uima/

Unstructured Information Management Architecture SDK. The UIMA SDK (Software Development Kit), is an all-JavaTM implementation of the UIMA framework, and it supports the implementation, description, composition, and deployment of UIMA components and applications. It also supports the developer with an Eclipse -based development environment that includes a set of tools and utilities for using UIMA.

Go you crazy Java dudes, go.

Curl

Laszlo

Flex

Create rich client apps with the DOM

Rich Cients ORG

---
  Your comment has too few characters per line (currently 5.7).

Actually, my Texas Instruments Voyage 200 graphing calculator uses a 14MHz Motorola MC68000 CPU with 384 KB RAM and 4 MB flash EEPROM. Last time I checked, the Mac PPC lineage traces back to that chip.

A hell of a lot more of its lineage traces back to the RS/6000, as per this history of the POWER family . (And the 68K was used for a lot more than calculators.)

(But, then again, the 8008 wasn't a calculator chip; they were thinking of the 4004.)

Uhm, additionally?

My post was about Linux & Windows... try keep on the topic I used, please, thanks.

BUT, there is RealBasic & I mentioned it in another reply here, it WILL write one codebase for Win32, Linux, & Macs. I have not tried it, and it is relatively an "infant" still, but imo?

A "baby hercules" really... especially for a 'write once, run everywhere in TRUE std. stand-alone .exe non-interpreted form" from a single codebase type of app on ALL 3 platforms.

Again - I haven't tried it, but like the sound of it already... especially imo, for the future of OS' being more "interoperable" & also having more common apps too!

By the by -

Linux folks & places like IBM as well, see the URL's below, HIGHLY AWARDED & RECEIVED KYLIX with great appreciation apparently:

(Also, @ tradeshows & such for it as well as mags for it)

Linux people have HIGHLY awarded & reviewed Kylix as an excellent tool!

See here in these:

http://www-128.ibm.com/developerworks/db2/library/ techarticle/0211swart/0211swart2.html

&

http://builder.com.com/5100-22-1064560.html

"Best Linux Development Tool: Borland Kylix 2
Borland wrapped up a double play by taking home our Best Linux Development Tool award for Kylix in yet another landslide. Kylix finished with 50 percent of the final vote. I was mildly surprised to see KDE Studio finish as first runner-up with 14.4 percent in our Linux category, slightly ahead of Metrowerks CodeWarrior for Linux with 12.4 percent. Its interesting to note that these three tools accounted for over 75 percent of the votes cast in this category, so obviously these companies are doing something right."

http://www.zdnetasia.com/builder/program/dev/0,390 45513,39163332,00.htm

http://www.programmersheaven.com/zone2/articles/ar ticle317.htm

"One of the potentially exciting things about Kylix is being able to write Apache web server shared modules."

http://sdmagazine.com/jolts/2002/year_01.htm

Tons more too, I could fill the page so... Heck, just search "Awards" and "Kylix" on GOOGLE, you will see I am not b.s.'ing you!

(32 pages worth of them) :)

APK

At work, a competition was formed. If you're interested, I can look up the rules we used.

Actually, IE renders entities this way... see this article on cross browser development: http://www-128.ibm.com/developerworks/web/library/ wa-ie2mozgd/#html_differences

But it looks like they have fixed it up now... spoil sports.

(Incidentally, I prefer NotesBuddy to the actual Sametime Connect app.)

The TrackPoint Mouse, anyone?

(This was productized in the '90s, with a blue nipple. CompUSA used to have racks of them.)

A difference I see between IBM and Microsoft, is that Big Blue seems to have at least done some truly innovative cutting edge stuff, I've yet to see something from Microsoft that makes me go "yeah that's realy impressive they came up with that". Some IBM researchers have actually received Nobel prizes, I've yet to see a Microsoft employee do that.
From http://www.research.ibm.com/know/top.html
- Copper Chip Technology
- Giant Magnetoresistive Head (GMR)
- Speech recognition technology
- Scalable parallel systems
- Token-ring networking
- High-temperature superconductivity (1987 Nobel prize)
- Fractals (Mandelbrot)
- Thin-film magnetic recording heads
- Scanning Tunneling Microscope (1986 Nobel prize)
- Formula Translation System (FORTRAN)
- Reduced Instruction Set Computer (RISC) architecture
- Relational database
- Magnetic disk storage
- One-transistor dynamic RAM (DRAM)

Disclaimer: 1) I am in no way affiliated with IBM, and know they abused their monopoly in the '80 too. I just think it's ridiculous to compare IBM and MS R&D results, just name 1 significant MS innovation.

2) I'm no big fan of software pattents in general, at least in their current form, because checking 20,000,000,000 pattents each time a program a line of code is impractical.

about their iSeries and such

http://www-1.ibm.com/servers/eserver/iseries/legen ds/index_flat.html

It also includes the "server" lost behind the wall. The reenactments are cute and somewhat based on "true" stories.

Another set of stories is at...
http://www.iseriesnetwork.com/nodeuk/ukarchive/ind ex.cfm?fuseaction=viewarticle&CO_ContentID=13885

You're confusing "privileged mode" AKA "kernel mode" with "in the kernel". Mach isn't a real microkernel, but it does support the ability to run software in privileged mode without it being "linked in the kernel".

No I'm not confusing the two. Accelerated graphics drivers have been buried so deeply into OSes nowadays that they are extremely hard to separate them from all internal APIs. If your driver allocates memory or accesses the PCI bus, it'll have to follow the license of the host OS code that does that (or disgustingly roll its own version of everything).

And Mach is *too* a microkernel. In fact its one of the most famous and influential microkernels ever written. The Mach project rewrote BSD into a microkernel. OSX is derived from an intermediate version. It is not a pure microkernel, just like NT is a hybrid. That's a pragmatic choice for performance. But what do I know about Mach, I've only read the Mach papers and taken graduate OS courses at the university where it was written (Carnegie Mellon), attended talks by some of the creators, and was a teaching assistant for an OS course at the same university.

Furthermore, whether or not something using a messaging protocol or RPC is "linked" or not hasn't really been answered from a licensing standpoint IMO. I'm sure the GNU Hurd people would argue that a lot of it is linking of some sort, and many others would disagree with them. At any rate, I know I wouldn't go there, especially not if I were a company. Apple is smart enough to avoid walking into a problem like that. What I said in the earlier post was in response to you saying that Apple could use Linux if it wanted to. Since (as you point out) Linux is not a microkernel, it would have to be linked, that's all that matters. Apple (smartly) went for BSD code. They are of course happy to use GPL code for userspace applications, but that's a different issue entirely.

Microsoft says lots of stuff, but it's their actions that count. And their only package that uses a significant amount of real open-source software uses both BSD and GPL components.

The code in their OS that's borrowed was taken from BSD. They are smart enough not to take GPL code. Even for userspace, as that would just bring bad press for their OS product.

Like HFS+?

Good point, that's one of their more significant contributions.

One operating system can't possibly fill all the niches. Linux is tied to a kernel design that works well for a certain class of problems,

Such as the class of problems falling between wristwatches and the world's top supercomputer? I'll settle for that.

... but it depends on Linus for continued development

Only as much as Apple depends on Steve Jobs. They are influential leaders, but not necessary. Look at LKML and you can watch yourself how things work. Someone could take over tomorrow if Linus could no longer work on the project (Andrew Morton or Alan Cox being likely successors).

... and Linus antipathy towards non-traditional kernels is legendary.

NT became successful after they dropped their devotion to a pure microkernel design. OSX has never been a pure microkernel from the start, being derived from an intermediate version of the Mach codebase. Linux has never dabbled with microkernel design since every successful OS has moved in the opposite direction. That's simply pragmatism, and all those OSes made the pragmatic choice.

In the last decade though, all OSes have adopted some of the simpler advantages of a microkernel design, such as modular drivers. If Hurd or DragonFlyBSD ever comes up with something compelling enough, I have no doubt Linux will adopt it. Up to now, there's been no reason to do so.

Have you looked at the the actual code? Or are you just going by "what

You're confusing "privileged mode" AKA "kernel mode" with "in the kernel". Mach isn't a real microkernel, but it does support the ability to run software in privileged mode without it being "linked in the kernel".

No I'm not confusing the two. Accelerated graphics drivers have been buried so deeply into OSes nowadays that they are extremely hard to separate them from all internal APIs. If your driver allocates memory or accesses the PCI bus, it'll have to follow the license of the host OS code that does that (or disgustingly roll its own version of everything).

And Mach is *too* a microkernel. In fact its one of the most famous and influential microkernels ever written. The Mach project rewrote BSD into a microkernel. OSX is derived from an intermediate version. It is not a pure microkernel, just like NT is a hybrid. That's a pragmatic choice for performance. But what do I know about Mach, I've only read the Mach papers and taken graduate OS courses at the university where it was written (Carnegie Mellon), attended talks by some of the creators, and was a teaching assistant for an OS course at the same university.

Furthermore, whether or not something using a messaging protocol or RPC is "linked" or not hasn't really been answered from a licensing standpoint IMO. I'm sure the GNU Hurd people would argue that a lot of it is linking of some sort, and many others would disagree with them. At any rate, I know I wouldn't go there, especially not if I were a company. Apple is smart enough to avoid walking into a problem like that. What I said in the earlier post was in response to you saying that Apple could use Linux if it wanted to. Since (as you point out) Linux is not a microkernel, it would have to be linked, that's all that matters. Apple (smartly) went for BSD code. They are of course happy to use GPL code for userspace applications, but that's a different issue entirely.

Microsoft says lots of stuff, but it's their actions that count. And their only package that uses a significant amount of real open-source software uses both BSD and GPL components.

The code in their OS that's borrowed was taken from BSD. They are smart enough not to take GPL code. Even for userspace, as that would just bring bad press for their OS product.

Like HFS+?

Good point, that's one of their more significant contributions.

One operating system can't possibly fill all the niches. Linux is tied to a kernel design that works well for a certain class of problems,

Such as the class of problems falling between wristwatches and the world's top supercomputer? I'll settle for that.

... but it depends on Linus for continued development

Only as much as Apple depends on Steve Jobs. They are influential leaders, but not necessary. Look at LKML and you can watch yourself how things work. Someone could take over tomorrow if Linus could no longer work on the project (Andrew Morton or Alan Cox being likely successors).

... and Linus antipathy towards non-traditional kernels is legendary.

NT became successful after they dropped their devotion to a pure microkernel design. OSX has never been a pure microkernel from the start, being derived from an intermediate version of the Mach codebase. Linux has never dabbled with microkernel design since every successful OS has moved in the opposite direction. That's simply pragmatism, and all those OSes made the pragmatic choice.

In the last decade though, all OSes have adopted some of the simpler advantages of a microkernel design, such as modular drivers. If Hurd or DragonFlyBSD ever comes up with something compelling enough, I have no doubt Linux will adopt it. Up to now, there's been no reason to do so.

Have you looked at the the actual code? Or are you just going by "what

A search on ibm.com does not give me a link to the document and neither does google. I did however find an IBM provided AMD vs XEON linpack-comparison benchmark ftp://ftp.software.ibm.com/eserver/benchmarks/wp_L inpack_072905.pdf, but first benchmark (that I can't locate) was better.

It clearly shows the advantages of the AMDs NUMA architecture and also other factors.

NUMA is also available on some enterprise level IBM XEON servers like the x440, x445 and x460 (or the equivelent systems from Fujitsu Siemens or NEC). One thing that is important on these servers is that you should balance each CEC with the same amount of memory or it will greatly affect performance. AMD's NUMA technology is not affected as much as XEON on this (as the mentioned paper shows).

Well, they already have IBM making some AMD64 servers ( http://www-132.ibm.com/webapp/wcs/stores/servlet/C ategoryDisplay?categoryId=2590406&storeId=1&catalo gId=-840&langId=-1 ). I can only hope IBM (don't know about others) had a broader line of AMD 64 based servers, but this is certainly a start.

I don't think you can kill it but you can optimise it, as the following research by IBM shows:

http://www-128.ibm.com/developerworks/power/librar y/pa-unrollav3/

It is true that Sun seems to have made some improvements. Weren't they supposed to fix ufs with Solaris 10? DOn't get me wrong. Solaris pays a lot of bills.

Solaris scale better then the others and runs well on 64-way multiprocessor systems

This Linux box scales to 64 processors. So does This AIX/Linux box.

Look at the IE addressbar (which is truncated) in this picture.

Makes you say "Hmmmmm..."

Here is an IBM site with some fun Linux stuff:
http://www-1.ibm.com/servers/eserver/linux/fun/

WARNING: The animations use evil flash magic.

Here is an IBM site with some fun Linux stuff:
http://www-1.ibm.com/servers/eserver/linux/fun/

A good example of this is the 'drawing an elephant' example given in this IBM document: pdf

2nd url was supposed to be this, oops!

Not to mention having the freedom to make superior machines without the Ivory Tower Syndrome.

Not to mention having the freedom to make superior machines without the Ivory Tower Syndrome.

IMNSHO, OpenFirmware is way too open.

PC industry is filled with crippled hardware. Only if people knew the quality of what they buy...

It is quite showing that OF is adopted only by Sun & Apple: both well know for quality harware with good record. Companies not afraid to respond to customers claims. And frequently responding.

Additionally, from rumors I have heard, EFI is designed madularly, and modules basicly can be any kind of binary blobs. So EFI will improve nothing on side of BIOS - it will remain closed source, and will quite quickly end-up being fix for every hole system might potentially have. Just like BIOS.

After all, all those magic stuff they promise, can be as easily implemented on top of modern BIOS. There is no technical problems as it is. *NO* *TECHNICAL* *PROBLEMS*. Many companies have source code for BIOS. Just sit down and do it. Just like IBM did with Slimline Open Firmware - stripped down version of OF, sufficient to boot Linux on JS20 blades.
http://www-128.ibm.com/developerworks/power/librar y/pa-dw-slof.html

P.S. From what I have heard about BIOS developement, it is low amount of flash which is problem: normally installed on PCs is only 64k. That's the problem - not BIOS itself. EFI I'm sure will mandate something like 1MB minimum of flash - after all the prices are now so much lower. But still as it stands: there is no un-solvable problem with BIOS.