Is it just me, or does the article have this backwards? Sure, I'm just nit-picking, but it seems that one would install such metal fibers to promote conduction and thus prevent the build-up of static charge. Consider conductive anti-stat floor mats, wrist straps, et cetera, which make this seem not-so unusual (except that it's built into the carpet).
No - it's not just you. I think they probably intended to say that the metal filaments helped to prevent a build up of static charge in any location. People get confused because static electricity is a fairly misleading name - really all they mean is that a potential difference builds up in a locality and is only slowly dispersed into the surroundings because of the high conductivity of the environment.
The GeForce 2MX is pretty damn sweet under xfree86, although I believe my shady motherboards APM was causing crashes.
I had instability nightmares for ages with my TNT2U on an AMD 750 chipset mobo, until I turned off the NVidia AGP support... After about six hours of coding (and maybe a couple of sessions of Quake III Arena:-) ) I suddenly realized that it hadn't crashed at all. Since then I have had only two crashes in four months, which is pitiful for a Linux machine but much better than a crash every two hours or so before the change.
Scan your/etc/X11/XF86Config-4 file for the Section "Screen"... EndSection. Add the following line
Options "NvAgp" "0"
It isn't guaranteed to fix every NVidia crash, but I've had reports from a few people that this fix has radically improved stability. Especially if you happen to have an Aureal Vortex soundcard in your system.
I have given up on Xfree 4.0.1 and gone back to 3.3.6 because the fonts appear to be completely messed up. At the same resolution, fonts, esp. truetype fonts, appear HUGE on xfree86. The fonts on netscape and Konqueror are especially bad. I don't know why this is, since I have setup Xfs to do the font serving and have merely set the FontPath to unix/:-1.
Sounds like your resolution settings are royally fubar'd. The renderer will try to guess the dots per inch from your monitor size and pixel resolution and will plot the fonts in an appropriate scale. It is sometimes useful to override this calculation - either
startx -dpi 100
or edit your/etc/X11/xdm/Xservers file to read
:0 local/usr/X11R6/bin/X -dpi 100
To be honest, I have no idea why you would want to use XFS alongside XFree86 4.0.1. The 4.0.x family includes rendering support for bitmap, Speedo, Type 1 and Truetype fonts in X and is therefore a lot easier to set up than using an external font server. I ran 'chkconfig xfs off' as soon as I installed 4.0 and I haven't looked back.
The mouse support providing DGA 1.0 in XFree86 4.0.1 was subtly broken and gave jumpy and unpredictable mouse behaviour which was most noticeable in Quake 3, among others. This has been fixed in the CVS Xfree86 tree for a few months now and I assume therefore it is fixed in 4.0.2.
Right-click blocking: Occasionally a banner ad will really start to tick you off, especially if it keeps coming from the same domain or URL. I'd like to be able to right click on it and choose "Block this URL", or "Block this (sub)domain" Or maybe just "Don't load any image that is placed right here."
Mozilla already allows you to block sites from loading images, in almost exactly this fashion. Right-click on the banner that you dislike, and choose 'Block image from loading' from the menu. This blocks all images from that site - fine for ads.doubleclick.net, etc. but it doesn't allow you to have a part-path yet - that may come. You can review which sites are currently blocked in the Image manager.
Firewall-like controls:I'd like to be able to tell Netscape/Mozilla to "block traffic from doubleclick.net", or whatever. I can do this if I mess with the firewall, but I'd rather leave it alone.
You can't block all traffic with Mozilla, but you can block the cookies from a site too. Enter the cookie manager and select the cookie you wish to remove. Check the 'Do not allow cookies from this site again' box and click 'Remove'. Of course this doesn't solve flash plugins or Javascript, but I'm fairly certain that as the need for these features grows, we'll see it added to the Mozilla codebase. At least with an open source project, adding this sort of functionality is possible.
I'm not about to state that Mozilla is the epitome of stability, but the number of crashes I have with build 2000121404 is none. But it's only been four days so far.
I also advocate running with Javascript off if you are visiting sites you suspect of useless popups.
It's getting harder to see where Corel is actually going these days. Still, the conspiracy theorists will have a field day linking this to the MS.NET support. But with sales of Corel Draw 9 not up where the company wanted them and increasing numbers of people happy with their current version of WordPerfect, it doesn't look like a particularly healthy revenue stream. MS has this problem too - people may be happy with the software they currently have (\{deity} help them) - but MS has the additional leverage of both new versions of the OS and the guarantee of more PC's being sold with MS software pre-installed on it.
Once again, the kernel tinkerers have provided us with another gem. This sort of experimentation with the core of an operating system is exactly the sort of thing that having the source code available encourages. While this implementation is probably not even safe to be considered beta software, it's good to see people playing with the boundaries.
To those who moan about code bloat in the kernel, they obviously haven't looked at how large the kernel sources are already - it's already huge. But this size doesn't matter - what matters is the code you choose to stick in your kernel - as almost all the peripheral code can be modularized or ignored, the core kernel loaded at boot time remains small and you can choose to ignore code you consider unnecessary or even dangerous, or put rarely used functions into modules to be loaded on a need-to-use basis, to be unloaded as soon as you are finished with that functionality.
So what possibilities does having the ORBit code in the kernel hold out in the future, for those who choose to make use of it? Faster, slicker handling of the CORBA transactions is the main one, and for those of us who might like to take the risk of destabilizing the system and experimenting with this stuff, we can.
It looks like they have latency problems with this code though - they should probably try integrating their code with Ingo Molnar's low latency kernel patches (average system latency under load down to 3ms - extremely nice for multimedia applications) and see whether their peak latency of ~1 second can be reduced. Oh - and by the way, those low-latency patches are not in the kernel sources, so you'll have to add them in yourself if you want to play with them. Much as you'll have to add the kORBit patches if you want to play with these either.
And finally, to those kernel hackers who do this sort of mad, mad, mad hacking, cheers!
Sure, AA might look beautiful in higher resolutions, but at a low resolution like 640x480 (or even 800x600), it looks like barf.
Total and utter *&^*&^^..:-)
Antialiasing improves the readability of a font at small sizes. That's why Acorn went to all the trouble of having it in Risc OS back in 1987- when your vertical resolution can be as low as 256 lines or less, keeping the fonts readable as the point size drops below 6 pts is impossible without anti-aliasing. They had this resolution because that was back in the days where people used their tellies as monitors.
Furthermore, some fonts were meant to be shown without anti-aliasing (MS Sans Serif, Times New Roman, and Arial in Windows; I'm sure there's some in X).
Thats because MS still hasn't got it's antialiasing working properly - hence MS 'font smoothing' is an appropriate title. Anti-aliasing is not just about blurring the edges - it is about increasing the apparent resolution of the text by using greyscales - the same way a truecolour phot has a higher apparent resolution than a black-and-white 2 colour image on the same display. Doing it right has a massive effect on the readability of the text on screen. Because MS's implementation doesn't cut it at small point sizes, they tweaked the truetype fonts to render more reliably to the screen instead.
There is only so much anti-aliasing will do to correct bad fonts.
Most of the bad fonts are bitmaps and aren't going to be affected by antialiasing unless they are rescaled bitmaps (excuse me, if you use rescaled bitmaps... I think I'm going to be sick...).
That said, there is no reason to restrict your font habits to truetype ones. Postscript Type 1 fonts can also be rendered well using antialiasing and can also give extremely good results. At the end of the day, the quality of the font on screen is limited by the quality of the hinting done to protect the important features of each glyph from being lost when rendered.
In windows, all of the true type fonts I use look great without anti-aliasing. If you want beautiful fonts in X windows use an X server that supports true type fonts.
Even at tiny point sizes? I doubt it. Antialiasing is good for increasing readability of fonts at all point sizes but especially for small fonts. Without antialiasing, small fonts become a muddle of pixels.
At some level this is nothing new. All professional sporting events in the US are all prefaced by a little statement that any recording or broadcasting of the event without the express written consent of whatever association runs the sport (NFL, NHL, NBA, etc.) is a crime.
I believe that the right of Fair Use still overrides these messages. If I want to post the latest score or write a review of my experiences at the last baseball game, they don't have a leg to stand on legally. Even if I post screenshots from the broadcast or possibly even short movie excerpts - the copyright laws allow limit reproduction. What they do have in their favour is a large selection of lawyers to send cease-and-desist demands around - does anyone here actually run an opensource sports news site? Have they had any grief?
Seriously, this isn't a minor plug-in a new backend sort of thing. DB2 is different with administration and the dialect of SQL is a bit different.
The fringes of the SQL may be different but the core SQL language is controlled by the SQL Language Council so any SQL which comes under the SQL '92 spec should move across without difficulty. It's the extras that cause the problems where the different vendors have different approaches to various problems.
Overall the level of industry use/acceptance of DB2 is much lower than Oracle, particularly the Unix/NT (UDB) version of DB2. DB2 is big in the mainframe world, but good luck finding new hires with DB2 on Unix experience. Mainframe experience doesn't automatically port over to Unix experience.
I'd like to know where you get your figures from. The level of industry use of DB2 is not 'much lower than Oracle', despite what Oracle's marketing department would have you believe. All those ads which say '96% of the Fortune 50 use Oracle' miss out the fact that most (all?) of the Fortune 50 use DB2. Most major companies use multiple database systems for various reasons. And an increasing proportion of those DB2 users are using UDB.
Should not use TeX as an intermediate format. TeX creates nice output, but it needs a complete rewrite to get rid of static data structures, and make the error messages human readable. Everyone who has used TeX for something useful has made major modifications to TeX to make it processes his/her particular document (changed 20 constants in a config file or recompiled the package). It amazes me that Donald Knuth has kept his guru status, when the second most known piece of work he has done is TeX... It certainly tells something about how excellent his most known work is (yes, it is excellent, this is not a flame).
Having written a 200-page technical document in LaTeX (I wish I'd known about LyX at the time!!) I'm afraid I don't share your view on TeX. Having put my document through the ringer, generating indices, tables of contents, multi cross references and three layers of sectioning, interspersed with multiple diagrams in multiple formats, I just did not have to fiddle with the base package at all. Armed with Leslie Lamports guide to LaTeX, there were no obscure error messages and any formatting decisions it made were logical, even if they weren't entirely what I intended. In the few cases where you get the infamous all-the-images-at-the-end-of-the-chapter problem, it just took a little rethinking and some coaxing in the LaTeX (not the C code) to get it where I wanted it.
And just in case I had a tweaked copy on the Solaris cluster I was using LaTeX on, I got identical results on my Linux box with the same files.
Cheers,
Toby Haynes
Even sterile GMOs can cross-pollinate successfully
on
Golden Rice
·
· Score: 2
What I liked about it was that the developers hadn't crippled the strain's ability to reproduce. Genetically-engineered wheat is generally crippled, forcing farmers to buy new seed from the company year after year.
But *if* this modified strain of rice should have some nasty side effect, it would be a really bad thing if it reproduced.
Yes - but just because GMO wheat can't produce viable seed from a reproduction standpoint (it's still edible) does not mean that it can't cross-pollinate with other plants, and pass it's attributes on to a plant with reproductive capability. There are a number of cases where a farmer whose 'natural' fields abutt a GMO crop have been sued by Monsanto (in another masterpiece of PR) for having wheat with the Monsanto-engineered attributes as a result of cross-pollination. I wait with interest to see what the courts make of this.
From a personal standpoint, I have little problem with GMO crops where the alterations would have arisen naturally sooner or later during normal reproduction. This is little more than weighting the dice, so to speak. Where you introduce a new protein into the DNA code to make it more resistant to, say, pesticides I have more concerns because this strays a lot further than evolution would have acheived simply by rolling the evolutionary die.
Cheers,
Toby Haynes
Observational Astronomers are already data miners
on
Creating The UniServer
·
· Score: 4
Your average astronomer is already a major data miner. From the Hubble Deep Field to the images taken in the back yard with a home-built CCD camera, much of modern observational astronomy is entirely built around being able to mine those images for correspondance, object attributes, clustering in either position, colour, or some other feature. Even with a basic catalogue built off one single wavelength plate will assign position, size, brightness, orientation, semi-major and semi-minor size, positional error, orientation error, brightness error, isophotal brightness, local background level and half-a-dozen other attributes to each object in the catalogue. There may be several thousand objects in a single frame. Making sense of this data set requires time, some ideas about what you are searching for and some luck.
All that said, you'd be missing a lot as an astronomer if all you looked at was optical images. Going to other images for the same area of sky, be it infra-red, radio, x-ray and so on, will give you a deeper insight into the likely environment of your object and also into any likely confusions due to multiple structures along the line of sight.
So having a vast data repository is important, and astronomers have had the tools to go and query multiple surveys at multiple wavelengths for several years. So there is nothing new here either from a data access point of view. The only really new thing in this proposal is to collate all the data together onto four super-mirrors and ensure that these supermirrors remain in sync, so if one system dies, it can be restored from the other mirrors without having to go back to tape backups.
The problem with eliminating physical ballots is that it leaves us with no recourse when an error occurs.
Which brings us back to the need for hard copy. This can be done by something as trivial as printing each vote cast out as a matrix which could be scanned by a bulk reader in the event of a recount or major system failure. Hey - you could even have it punch holes in card if you feel nostalgic:-)
Yeah, I've been using M18 (or nightly builds of it, actually) as my primary browser on Linux for a while now. It still doesn't have https support so I have to use Njetscape 4.7 to access ecommerce sites, but that's about the only thing I still use NS for.
Assumming you are running on either Windows, Solaris or Linux, you can 'Install PSM' from the 'Debug' Menu at the top of the screen. Scroll down the web page and click the appropriate button for your OS (or load the package in manually for Solaris). If all goes well, you should see the package load in and your should get a successful XPInstall message. It would be nice to see some more OS's supported - at least MacOS is pending and a BSD-compatible version and some for other Unix platforms and architectures would be nice. Maybe there is room for an OpenSource PSM project.
PSM is good enough that I've successfully ordered plane tickets using it, and can quite comfortably browse Sourceforge in SSL mode.
I've never really been a java fan, but if you get the chance, you should really check out the most recent version of the jre (1.3). It runs about 3x faster than any previous version.
Cool.
Yes.
Does it run on Mac? No?
No - no Mac version at the moment. It will be along presently.
Supported in major browsers? No?
Supported by Netscape 6 and Mozilla M18+ - runs nicely on both Windows NT and Linux.
I think I'll stay "cross-platform" and stick with 1.1.7.
I'll concur on the need for Java 2 RE to spread to more platforms - I assume there is a Sparc version in there somewhere, and MacOS will be there soon I believe. Not too shabby.
I tried out IE for Solaris. It's missing a lot of the features of the Win32 version. If you install it, you'll also see that Microsoft doesn't quite understand this "UNIX" thing... half of the install goes in your home directory (rather than somewhere in/usr), but with the permissions set so that only root can execute the browser.
I've tried IE on Solaris. I can't think of another single application which had such resource usage of the system. The UltraSparc I was running on felt like a 486 when browsing with IE, and top constantly reported massive load on the system while IE was running. I kept switching back and forth between Netscape 4.5 and IE on the system trying to get a feel for the differences, and IE was taking about twice the time to do almost anything, with the exception of table rendering, where it was faster than Netscape on massive tables, and of course on resizing the page where NS 4.5 reloads the page.
Made me wonder just how far through the Windows OS the roots of IE actually went. The only thing that could account for such usage in a port would be subsidiary support services running to keep the IE browser happy in a Unix environment.
Windows implementation of Antialiasing (known as Font Smoothing in the Windows world) is a long way away from the ideal Nirvana of text presentation. Antialiasing is really a problem with small text sizes - all those serifs and lines close together get confused when you try to render a vector object onto a grid with too few sampling points. As Windows only smooths the larger font sizes by default, this makes it a little irrelevent for text viewing. Nyquist would tell you more.
Therefore an advanced font renderer will help the eye perceive the real shape of the text better by shading the text with different shades between the text by understanding what proportion of the curve of the letter lies inside the pixel. This is particularly important when you consider sub pixel alignment both horizontally and vertically - with large passages of text with real micro alignment, this makes a huge difference to the readability of the text - enough that helvetica in a 5 pixel high font is vageuly intelligible. Speaking as someone who tend to use a 6x13 font for most things on a 21 inch monitor, antialiasing of the text in Type1 or truetype vector formats would be a huge step forward.
Cheers,
Toby Haynes
She's not breaking the sound barrier in thin air.
on
Sub-Orbital Skydiving
·
· Score: 3
Wow, that's fast... I was under the impression that breaking the sound barrier was rather stressful for whatever does it, since the sound waves it creates can't get out of its way before it plows into them. I guess a lone parachuter may not be making much sound, especially until she hits the atmosphere, but once she does reach air it seems to me that the amount of drag on her body would amplify greatly due to the high speeds and the effect of the sound barrier.
The speed Mach 1.5 is a little misleading - in the air she'll be travelling through, she won't be breaking the speed of sound, although she may reach around 1000 miles/hour. In thin air, the speed of sound is much higher than at sea level. She will actually decelerate as she falls into the thicker lower atmosphere, so at no time will she be going fast than the speed of sound in the air she is travelling through. The main danger as far as I can see is difficulty in preventing a spin in a thin atmosphere - there is little air resistance to allow you to stop rotation motion. Still, for an experienced sky diver, this shouldn't be an insurmountable problem.
sorry, my bad. It is free, but incompatible with the GPL.
http://www.gnu.org/philosophy/li cen se-list.html
Is it? I don't see it on that list. I see the IBM Public License listed as GPL incompatible, but that doesn't mean that the IBM Open Source license is the same.
I wonder if it'd be easier/faster to 'chattr +i' certain critical files like/bin/login and then add logging code to the appropriate syscall to warn you when somebody changes it back.
Why allow anyone to change this back? Set the immutable attribute (+i) on anything that won't ever change, make your log records permanent with 'chattr +a/var/log/messages' so that the logs can't be editted, just appended to, and then install lids and set CAP_LINUX_IMMUTABLE to remove the ability to change these attributes on this system under this kernel. If you need to change things over, you'll need a second kernel image elsewhere for administration purposes (i.e. on floppy) but your key system will remain inviolate.
Are TPC benchmarks useful? Really when you think about it it's more a benchmark of how well the clustering and hardware performed during the test. The top spot which Microsoft holds on a Compaq machine is a 192 processor Xeon cluster with a huge number of dual processor clients etc.
These benchmarks are useful - they give the performance teams at IBM something to aim at and keep us on our toes in development:-)
On a more serious note, they give you a good idea of how these databases scale when you are talking about SMP/MPP systems. Oracle suffers in large clusters of machines because they have a shared disk array system, so there is always an overhead keeping all the parts of the cluster in sync. Both DB2 UDB and SQL Server have a shared nothing architecture, which means you have to carry copies of certain data structures around the network, but you only take it where it's needed. This wins when you get to 8 or more nodes in a cluster and gets better as you add more. Having said that, it's interesting to compare that SQL Server TPCC benchmark done on a 192 processor cluster scoring aroudn 520000 transactions against the 128 processor DB2 UDB benchmark (same CPU type) scoring 440000. Take your pick as to which is the faster database (hint DB2 UDB scales nearly linearly as you add cluster nodes).
Not that I like a bit of shameless self promotion;-)
Not allow benchmarks to be published without permission seems to be the standard for the database companies. I know Microsoft, Oracle, Sybase, and Informix all do this.
Not true for DB2 from IBM. Okay - so I work on DB2 for IBM Canada but I would be seriously worried if we prohibited people from running, testing and comparing our software against the competition. Oracle's license makes a nonsense of the already restrictive license agreements that most commercial software comes with. But ask yourself this - why would a software company not want you to run benchmarks on their software?
Are you kidding? You calling this BARS? Wait until some students get the wrong end of the stick and take this into the pub with them?
Auto-readouts of which beers are on special? Which barman/barmaids to chat up? GPS-assisted washroom finding facilities? Auto-pilot home after 15 pints of Abbot Ale? (Actually most students seem to manage this without assistance).
Slashdot Mirror
Is it just me, or does the article have this backwards? Sure, I'm just nit-picking, but it seems that one would install such metal fibers to promote conduction and thus prevent the build-up of static charge. Consider conductive anti-stat floor mats, wrist straps, et cetera, which make this seem not-so unusual (except that it's built into the carpet).
No - it's not just you. I think they probably intended to say that the metal filaments helped to prevent a build up of static charge in any location. People get confused because static electricity is a fairly misleading name - really all they mean is that a potential difference builds up in a locality and is only slowly dispersed into the surroundings because of the high conductivity of the environment.
Cheers,
Toby Haynes
The GeForce 2MX is pretty damn sweet under xfree86, although I believe my shady motherboards APM was causing crashes.
I had instability nightmares for ages with my TNT2U on an AMD 750 chipset mobo, until I turned off the NVidia AGP support... After about six hours of coding (and maybe a couple of sessions of Quake III Arena :-) ) I suddenly realized that it hadn't crashed at all. Since then I have had only two crashes in four months, which is pitiful for a Linux machine but much better than a crash every two hours or so before the change.
Scan your /etc/X11/XF86Config-4 file for the Section "Screen" ... EndSection. Add the following line
It isn't guaranteed to fix every NVidia crash, but I've had reports from a few people that this fix has radically improved stability. Especially if you happen to have an Aureal Vortex soundcard in your system.
Cheers,
Toby Haynes
I have given up on Xfree 4.0.1 and gone back to 3.3.6 because the fonts appear to be completely messed up. At the same resolution, fonts, esp. truetype fonts, appear HUGE on xfree86. The fonts on netscape and Konqueror are especially bad. I don't know why this is, since I have setup Xfs to do the font serving and have merely set the FontPath to unix/:-1.
Sounds like your resolution settings are royally fubar'd. The renderer will try to guess the dots per inch from your monitor size and pixel resolution and will plot the fonts in an appropriate scale. It is sometimes useful to override this calculation - either
or edit your /etc/X11/xdm/Xservers file to read
To be honest, I have no idea why you would want to use XFS alongside XFree86 4.0.1. The 4.0.x family includes rendering support for bitmap, Speedo, Type 1 and Truetype fonts in X and is therefore a lot easier to set up than using an external font server. I ran 'chkconfig xfs off' as soon as I installed 4.0 and I haven't looked back.
Cheers,
Toby Haynes
The mouse support providing DGA 1.0 in XFree86 4.0.1 was subtly broken and gave jumpy and unpredictable mouse behaviour which was most noticeable in Quake 3, among others. This has been fixed in the CVS Xfree86 tree for a few months now and I assume therefore it is fixed in 4.0.2.
Cheers,
Toby Haynes
Right-click blocking: Occasionally a banner ad will really start to tick you off, especially if it keeps coming from the same domain or URL. I'd like to be able to right click on it and choose "Block this URL", or "Block this (sub)domain" Or maybe just "Don't load any image that is placed right here."
Mozilla already allows you to block sites from loading images, in almost exactly this fashion. Right-click on the banner that you dislike, and choose 'Block image from loading' from the menu. This blocks all images from that site - fine for ads.doubleclick.net, etc. but it doesn't allow you to have a part-path yet - that may come. You can review which sites are currently blocked in the Image manager.
Firewall-like controls:I'd like to be able to tell Netscape/Mozilla to "block traffic from doubleclick.net", or whatever. I can do this if I mess with the firewall, but I'd rather leave it alone.
You can't block all traffic with Mozilla, but you can block the cookies from a site too. Enter the cookie manager and select the cookie you wish to remove. Check the 'Do not allow cookies from this site again' box and click 'Remove'. Of course this doesn't solve flash plugins or Javascript, but I'm fairly certain that as the need for these features grows, we'll see it added to the Mozilla codebase. At least with an open source project, adding this sort of functionality is possible.
I'm not about to state that Mozilla is the epitome of stability, but the number of crashes I have with build 2000121404 is none. But it's only been four days so far.
I also advocate running with Javascript off if you are visiting sites you suspect of useless popups.
Cheers,
Toby Haynes
... hasn't a leg to stand on :-)
It's getting harder to see where Corel is actually going these days. Still, the conspiracy theorists will have a field day linking this to the MS .NET support. But with sales of Corel Draw 9 not up where the company wanted them and increasing numbers of people happy with their current version of WordPerfect, it doesn't look like a particularly healthy revenue stream. MS has this problem too - people may be happy with the software they currently have (\{deity} help them) - but MS has the additional leverage of both new versions of the OS and the guarantee of more PC's being sold with MS software pre-installed on it.
Cheers,
Toby Haynes
Once again, the kernel tinkerers have provided us with another gem. This sort of experimentation with the core of an operating system is exactly the sort of thing that having the source code available encourages. While this implementation is probably not even safe to be considered beta software, it's good to see people playing with the boundaries.
To those who moan about code bloat in the kernel, they obviously haven't looked at how large the kernel sources are already - it's already huge. But this size doesn't matter - what matters is the code you choose to stick in your kernel - as almost all the peripheral code can be modularized or ignored, the core kernel loaded at boot time remains small and you can choose to ignore code you consider unnecessary or even dangerous, or put rarely used functions into modules to be loaded on a need-to-use basis, to be unloaded as soon as you are finished with that functionality.
So what possibilities does having the ORBit code in the kernel hold out in the future, for those who choose to make use of it? Faster, slicker handling of the CORBA transactions is the main one, and for those of us who might like to take the risk of destabilizing the system and experimenting with this stuff, we can.
It looks like they have latency problems with this code though - they should probably try integrating their code with Ingo Molnar's low latency kernel patches (average system latency under load down to 3ms - extremely nice for multimedia applications) and see whether their peak latency of ~1 second can be reduced. Oh - and by the way, those low-latency patches are not in the kernel sources, so you'll have to add them in yourself if you want to play with them. Much as you'll have to add the kORBit patches if you want to play with these either.
And finally, to those kernel hackers who do this sort of mad, mad, mad hacking, cheers!
Toby Haynes
Sure, AA might look beautiful in higher resolutions, but at a low resolution like 640x480 (or even 800x600), it looks like barf.
Total and utter *&^*&^^.. :-)
Antialiasing improves the readability of a font at small sizes. That's why Acorn went to all the trouble of having it in Risc OS back in 1987- when your vertical resolution can be as low as 256 lines or less, keeping the fonts readable as the point size drops below 6 pts is impossible without anti-aliasing. They had this resolution because that was back in the days where people used their tellies as monitors.
Furthermore, some fonts were meant to be shown without anti-aliasing (MS Sans Serif, Times New Roman, and Arial in Windows; I'm sure there's some in X).
Thats because MS still hasn't got it's antialiasing working properly - hence MS 'font smoothing' is an appropriate title. Anti-aliasing is not just about blurring the edges - it is about increasing the apparent resolution of the text by using greyscales - the same way a truecolour phot has a higher apparent resolution than a black-and-white 2 colour image on the same display. Doing it right has a massive effect on the readability of the text on screen. Because MS's implementation doesn't cut it at small point sizes, they tweaked the truetype fonts to render more reliably to the screen instead.
Cheers,
Toby Haynes
There is only so much anti-aliasing will do to correct bad fonts.
Most of the bad fonts are bitmaps and aren't going to be affected by antialiasing unless they are rescaled bitmaps (excuse me, if you use rescaled bitmaps... I think I'm going to be sick...).
That said, there is no reason to restrict your font habits to truetype ones. Postscript Type 1 fonts can also be rendered well using antialiasing and can also give extremely good results. At the end of the day, the quality of the font on screen is limited by the quality of the hinting done to protect the important features of each glyph from being lost when rendered.
In windows, all of the true type fonts I use look great without anti-aliasing. If you want beautiful fonts in X windows use an X server that supports true type fonts.
Even at tiny point sizes? I doubt it. Antialiasing is good for increasing readability of fonts at all point sizes but especially for small fonts. Without antialiasing, small fonts become a muddle of pixels.
Cheers,
Toby Haynes
At some level this is nothing new. All professional sporting events in the US are all prefaced by a little statement that any recording or broadcasting of the event without the express written consent of whatever association runs the sport (NFL, NHL, NBA, etc.) is a crime.
I believe that the right of Fair Use still overrides these messages. If I want to post the latest score or write a review of my experiences at the last baseball game, they don't have a leg to stand on legally. Even if I post screenshots from the broadcast or possibly even short movie excerpts - the copyright laws allow limit reproduction. What they do have in their favour is a large selection of lawyers to send cease-and-desist demands around - does anyone here actually run an opensource sports news site? Have they had any grief?
Cheers,
Toby Haynes
Seriously, this isn't a minor plug-in a new backend sort of thing. DB2 is different with administration and the dialect of SQL is a bit different.
The fringes of the SQL may be different but the core SQL language is controlled by the SQL Language Council so any SQL which comes under the SQL '92 spec should move across without difficulty. It's the extras that cause the problems where the different vendors have different approaches to various problems.
Overall the level of industry use/acceptance of DB2 is much lower than Oracle, particularly the Unix/NT (UDB) version of DB2. DB2 is big in the mainframe world, but good luck finding new hires with DB2 on Unix experience. Mainframe experience doesn't automatically port over to Unix experience.
I'd like to know where you get your figures from. The level of industry use of DB2 is not 'much lower than Oracle', despite what Oracle's marketing department would have you believe. All those ads which say '96% of the Fortune 50 use Oracle' miss out the fact that most (all?) of the Fortune 50 use DB2. Most major companies use multiple database systems for various reasons. And an increasing proportion of those DB2 users are using UDB.
Cheers,
Toby Haynes
DB2 UDB developer
Should not use TeX as an intermediate format. TeX creates nice output, but it needs a complete rewrite to get rid of static data structures, and make the error messages human readable. Everyone who has used TeX for something useful has made major modifications to TeX to make it processes his/her particular document (changed 20 constants in a config file or recompiled the package). It amazes me that Donald Knuth has kept his guru status, when the second most known piece of work he has done is TeX ... It certainly tells something about how excellent his most known work is (yes, it is excellent, this is not a flame).
Having written a 200-page technical document in LaTeX (I wish I'd known about LyX at the time!!) I'm afraid I don't share your view on TeX. Having put my document through the ringer, generating indices, tables of contents, multi cross references and three layers of sectioning, interspersed with multiple diagrams in multiple formats, I just did not have to fiddle with the base package at all. Armed with Leslie Lamports guide to LaTeX, there were no obscure error messages and any formatting decisions it made were logical, even if they weren't entirely what I intended. In the few cases where you get the infamous all-the-images-at-the-end-of-the-chapter problem, it just took a little rethinking and some coaxing in the LaTeX (not the C code) to get it where I wanted it.
And just in case I had a tweaked copy on the Solaris cluster I was using LaTeX on, I got identical results on my Linux box with the same files.
Cheers,
Toby Haynes
What I liked about it was that the developers hadn't crippled the strain's ability to reproduce. Genetically-engineered wheat is generally crippled, forcing farmers to buy new seed from the company year after year.
But *if* this modified strain of rice should have some nasty side effect, it would be a really bad thing if it reproduced.
Yes - but just because GMO wheat can't produce viable seed from a reproduction standpoint (it's still edible) does not mean that it can't cross-pollinate with other plants, and pass it's attributes on to a plant with reproductive capability. There are a number of cases where a farmer whose 'natural' fields abutt a GMO crop have been sued by Monsanto (in another masterpiece of PR) for having wheat with the Monsanto-engineered attributes as a result of cross-pollination. I wait with interest to see what the courts make of this.
From a personal standpoint, I have little problem with GMO crops where the alterations would have arisen naturally sooner or later during normal reproduction. This is little more than weighting the dice, so to speak. Where you introduce a new protein into the DNA code to make it more resistant to, say, pesticides I have more concerns because this strays a lot further than evolution would have acheived simply by rolling the evolutionary die.
Cheers,
Toby Haynes
Your average astronomer is already a major data miner. From the Hubble Deep Field to the images taken in the back yard with a home-built CCD camera, much of modern observational astronomy is entirely built around being able to mine those images for correspondance, object attributes, clustering in either position, colour, or some other feature. Even with a basic catalogue built off one single wavelength plate will assign position, size, brightness, orientation, semi-major and semi-minor size, positional error, orientation error, brightness error, isophotal brightness, local background level and half-a-dozen other attributes to each object in the catalogue. There may be several thousand objects in a single frame. Making sense of this data set requires time, some ideas about what you are searching for and some luck.
All that said, you'd be missing a lot as an astronomer if all you looked at was optical images. Going to other images for the same area of sky, be it infra-red, radio, x-ray and so on, will give you a deeper insight into the likely environment of your object and also into any likely confusions due to multiple structures along the line of sight.
So having a vast data repository is important, and astronomers have had the tools to go and query multiple surveys at multiple wavelengths for several years. So there is nothing new here either from a data access point of view. The only really new thing in this proposal is to collate all the data together onto four super-mirrors and ensure that these supermirrors remain in sync, so if one system dies, it can be restored from the other mirrors without having to go back to tape backups.
Cheers,
Toby Haynes
The problem with eliminating physical ballots is that it leaves us with no recourse when an error occurs.
Which brings us back to the need for hard copy. This can be done by something as trivial as printing each vote cast out as a matrix which could be scanned by a bulk reader in the event of a recount or major system failure. Hey - you could even have it punch holes in card if you feel nostalgic :-)
Cheers,
Toby Haynes
Yeah, I've been using M18 (or nightly builds of it, actually) as my primary browser on Linux for a while now. It still doesn't have https support so I have to use Njetscape 4.7 to access ecommerce sites, but that's about the only thing I still use NS for.
Assumming you are running on either Windows, Solaris or Linux, you can 'Install PSM' from the 'Debug' Menu at the top of the screen. Scroll down the web page and click the appropriate button for your OS (or load the package in manually for Solaris). If all goes well, you should see the package load in and your should get a successful XPInstall message. It would be nice to see some more OS's supported - at least MacOS is pending and a BSD-compatible version and some for other Unix platforms and architectures would be nice. Maybe there is room for an OpenSource PSM project.
PSM is good enough that I've successfully ordered plane tickets using it, and can quite comfortably browse Sourceforge in SSL mode.
Cheers,
Toby Haynes
I've never really been a java fan, but if you get the chance, you should really check out the most recent version of the jre (1.3). It runs about 3x faster than any previous version.
Cool.
Yes.
Does it run on Mac? No?
No - no Mac version at the moment. It will be along presently.
Supported in major browsers? No?
Supported by Netscape 6 and Mozilla M18+ - runs nicely on both Windows NT and Linux.
I think I'll stay "cross-platform" and stick with 1.1.7.
I'll concur on the need for Java 2 RE to spread to more platforms - I assume there is a Sparc version in there somewhere, and MacOS will be there soon I believe. Not too shabby.
Cheers,
Toby Haynes
I tried out IE for Solaris. It's missing a lot of the features of the Win32 version. If you install it, you'll also see that Microsoft doesn't quite understand this "UNIX" thing... half of the install goes in your home directory (rather than somewhere in /usr), but with the permissions set so that only root can execute the browser.
I've tried IE on Solaris. I can't think of another single application which had such resource usage of the system. The UltraSparc I was running on felt like a 486 when browsing with IE, and top constantly reported massive load on the system while IE was running. I kept switching back and forth between Netscape 4.5 and IE on the system trying to get a feel for the differences, and IE was taking about twice the time to do almost anything, with the exception of table rendering, where it was faster than Netscape on massive tables, and of course on resizing the page where NS 4.5 reloads the page.
Made me wonder just how far through the Windows OS the roots of IE actually went. The only thing that could account for such usage in a port would be subsidiary support services running to keep the IE browser happy in a Unix environment.
Cheers,
Toby Haynes
Because Windows has had it for 5 years.
And RiscOS had it in 1987. Your point?
Windows implementation of Antialiasing (known as Font Smoothing in the Windows world) is a long way away from the ideal Nirvana of text presentation. Antialiasing is really a problem with small text sizes - all those serifs and lines close together get confused when you try to render a vector object onto a grid with too few sampling points. As Windows only smooths the larger font sizes by default, this makes it a little irrelevent for text viewing. Nyquist would tell you more.
Therefore an advanced font renderer will help the eye perceive the real shape of the text better by shading the text with different shades between the text by understanding what proportion of the curve of the letter lies inside the pixel. This is particularly important when you consider sub pixel alignment both horizontally and vertically - with large passages of text with real micro alignment, this makes a huge difference to the readability of the text - enough that helvetica in a 5 pixel high font is vageuly intelligible. Speaking as someone who tend to use a 6x13 font for most things on a 21 inch monitor, antialiasing of the text in Type1 or truetype vector formats would be a huge step forward.
Cheers,
Toby Haynes
Wow, that's fast... I was under the impression that breaking the sound barrier was rather stressful for whatever does it, since the sound waves it creates can't get out of its way before it plows into them. I guess a lone parachuter may not be making much sound, especially until she hits the atmosphere, but once she does reach air it seems to me that the amount of drag on her body would amplify greatly due to the high speeds and the effect of the sound barrier.
The speed Mach 1.5 is a little misleading - in the air she'll be travelling through, she won't be breaking the speed of sound, although she may reach around 1000 miles/hour. In thin air, the speed of sound is much higher than at sea level. She will actually decelerate as she falls into the thicker lower atmosphere, so at no time will she be going fast than the speed of sound in the air she is travelling through. The main danger as far as I can see is difficulty in preventing a spin in a thin atmosphere - there is little air resistance to allow you to stop rotation motion. Still, for an experienced sky diver, this shouldn't be an insurmountable problem.
Cheers,
Toby Haynes
sorry, my bad. It is free, but incompatible with the GPL. http://www.gnu.org/philosophy/li cen se-list.html
Is it? I don't see it on that list. I see the IBM Public License listed as GPL incompatible, but that doesn't mean that the IBM Open Source license is the same.
Cheers,
Toby Haynes
I wonder if it'd be easier/faster to 'chattr +i' certain critical files like /bin/login and then add logging code to the appropriate syscall to warn you when somebody changes it back.
Why allow anyone to change this back? Set the immutable attribute (+i) on anything that won't ever change, make your log records permanent with 'chattr +a /var/log/messages' so that the logs can't be editted, just appended to, and then install lids and set CAP_LINUX_IMMUTABLE to remove the ability to change these attributes on this system under this kernel. If you need to change things over, you'll need a second kernel image elsewhere for administration purposes (i.e. on floppy) but your key system will remain inviolate.
Cheers,
Toby Haynes
Are TPC benchmarks useful? Really when you think about it it's more a benchmark of how well the clustering and hardware performed during the test. The top spot which Microsoft holds on a Compaq machine is a 192 processor Xeon cluster with a huge number of dual processor clients etc.
These benchmarks are useful - they give the performance teams at IBM something to aim at and keep us on our toes in development :-)
On a more serious note, they give you a good idea of how these databases scale when you are talking about SMP/MPP systems. Oracle suffers in large clusters of machines because they have a shared disk array system, so there is always an overhead keeping all the parts of the cluster in sync. Both DB2 UDB and SQL Server have a shared nothing architecture, which means you have to carry copies of certain data structures around the network, but you only take it where it's needed. This wins when you get to 8 or more nodes in a cluster and gets better as you add more. Having said that, it's interesting to compare that SQL Server TPCC benchmark done on a 192 processor cluster scoring aroudn 520000 transactions against the 128 processor DB2 UDB benchmark (same CPU type) scoring 440000. Take your pick as to which is the faster database (hint DB2 UDB scales nearly linearly as you add cluster nodes).
Not that I like a bit of shameless self promotion ;-)
Cheers,
Toby Haynes
Not allow benchmarks to be published without permission seems to be the standard for the database companies. I know Microsoft, Oracle, Sybase, and Informix all do this.
Not true for DB2 from IBM. Okay - so I work on DB2 for IBM Canada but I would be seriously worried if we prohibited people from running, testing and comparing our software against the competition. Oracle's license makes a nonsense of the already restrictive license agreements that most commercial software comes with. But ask yourself this - why would a software company not want you to run benchmarks on their software?
Cheers,
Toby Haynes
Are you kidding? You calling this BARS? Wait until some students get the wrong end of the stick and take this into the pub with them?
Auto-readouts of which beers are on special? Which barman/barmaids to chat up? GPS-assisted washroom finding facilities? Auto-pilot home after 15 pints of Abbot Ale? (Actually most students seem to manage this without assistance).
Cheers,
Toby Haynes