A bit over-idealistic, but worth aiming towards even if you don't achieve 100% non-URI-breakage in practice.
I feel that search engines should slightly penalize sites that have a history of breaking links or making them redirect to a completely irrelevant page: partly because there is just less chance that the link you follow from the search engine will have the content you want, and partly because even if you do get to a correct page, its usefulness as a bookmark or a link from your own dcuments is reduced.
They make no claim to assess whether music is 'good' or 'crap'. All they try to predict is whether it will sell.
If 'what's hot now' stops selling in large numbers then the algorithm will be adjusted - presumably they keep feeding in the latest songs and their sales volumes.
Yeah I wasn't talking about what Javascript _could_ be used for an an ideal world where all web page authors are non-cracksmoking, but about what it has been used for in practice, which is 99% crud. And that has been the case right back to 1994.
So what's an example of something useful (not advertising or moronic animations) that can be done in Javascript but not done using mid-1990s HTML? Form validation is one example, that would be the one per cent opposed to the 99% of crap I mentioned. But anything apart from that?
I've never understood why sites such as Slashdot want to use cookies for authorization... what's wrong with the standard username/password dialogue box that the browser is able to pop up and which servers like Apache easily support? That means a consistent user interface for logging in across sites, and if the user presses Cancel then he's taken to a sign-up page. The New York Times site seems to be one of the few that does this right.
As far as I recall 99% of Javascript has always been crap and useless and a security hole right from the day it was introduced; the part that has only happened recently is browsers (eg Mozilla Firebird) being sensible enough to block popups and other egregious behaviour. But the reason for that is that during the 1990s there was a too-cosy between the big browser makers (Netscape, Microsoft) and the advertisers that made money (or at least had a good stock price while losing money) through crappy 'portals'.
Ask Igor seems to be an implementation of delta debugging. You can use the delta program to implement this on your own system. You choose a test program (or 'harness') and an input that causes the harness to exit with success; for example your harness might run some executable and test to see if it segfaults - if so success. Then you give an initial input that passes the test (eg causes the segfault). Delta chops out lines of text to find a minimal (or at least 1-minimal, see the website) test case that passes the test (causes the segfault).
This is slightly different to Ask Igor, which takes two different files and finds the important difference between them. But similar in spirit (if much simpler). Apparently the Ask Igor code will be made available for download after it has been used 1000 times from the website.
What exactly causes this phenomenon to be specific to Ipods? It could happen equally well with walkmen or anything having a headphone socket. Is it just because Ipod users are more likely to blog about it and to read others' blogs, causing the practice to take off?
BTW - Ipods are quite expensive, this sounds like a good way to mug people for them.
A separate paper trail is good, but it would seem simpler to just make the main voting record be paper. There are plenty of automated multiple-choice tests where you fill in a box with a pencil and the machine reads your answers. This could be done for voting and realize just the same cost savings as touchscreen voting, but with a built-in paper audit trail. Furthermore, the ballots could be processed by two systems from competing vendors to make sure the results agree. As a last resort they can be counted by hand.
I just don't understand why touch screens are so favoured compared to paper.
So which came first: the Cornell project, or the Fedora project at the University of Hawaii? If Cornell's project was first, why haven't they complained about the name clash until now?
If you are running the software on your own laptop, then you don't need to care about 'vulnerabilities'. It's your laptop so you can run what you like on it anyway! For example, run the exam software in a VMware virtual machine or under a debugger. If you demonstrate this point to the exam organizers I'm sure they will rethink their plan.
To do online exams you need to control the PCs being used, as done with Lexis.
So let me get this straight. NASA has installed a big new Linux-based supercomputer, consuming only 5.2kW of power per rack, that has broken previous transaction processing benchmarks running Oracle 10g.
It should be possible to apply the same technique to other performance-tuning things like buffer sizes, or even choosing between two different algorithms for the same task. Perhaps a program could have a single header file performance.h containing// buffer size for reading
const size_t read_bufsiz = 1000;// amount of new memory to allocate in a chunk
const size_t chunk_size = 5000;// threshold for switching from quicksort to// merge sort when you have this few elements//
const unsigned int merge_sort_thresh = 10;
and then these values would be tuned at the same time as compiler flags.
Sorry, I don't know where ISOs of RHEL are available for download. (That in itself wouldn't stop it being free - for example, the FSF does not provide ISO images of GNU software for download but they will be happy to sell you a physical CD for $5000.)
Doesn't the above apply only if you sign a support contract with Red Hat? Of course you can sign a contract with another company for any stupid conditions you want, but you could also get the software without restrictions and without support (which puts it in the same category as, say, Slackware). If Red Hat won't sell you support except under draconian rules, don't buy it from them.
Yes, Netscape made a scripting language and Sun was happy to stick a meaningless 'Java' label on it although it has nothing to do with Java. Sun agreed to let Netscape use the trademark.
In this case the obvious cure is to render the 'HTML' to plain text first and then do spam-checking on that. Of course if you use a lame mail reader that really wants to display the lovely red colours and FONT SIZE="+9" then you still have a mismatch between what is checked and what is displayed, but not such a big one.
Why not Linux's remote block device? It might perform better than NFS because (assuming read-only access to the block device on the server, with a smaller device for/var) blocks can be cached on the client. (OTOH, a clever NFS implementation could cache bits of filesystem if you promised it the data was not going to change.)
So is there an authentication system for NFS that doesn't trust the client PC? One that is secure against plugging in your own laptop and setting it to have the same IP address and MAC address as an existing machine (which you unplug at the same time)?
I believe that in Sun's current worldview Java is a meaningless trademark to be slapped onto anything: hence Java Desktop. We should have seen the warning signs early with Javascript.
This is Euro-legislation. It's the result of lobbying in Brussels. Then a vote by ministers from all EU countries to adopt the directive - you know, France, Belgium, Germany, all those countries that are far up Bush's arse.
According to ARIN there are about 47 class A networks in the reserved status (search ARIN for OrgName "Internet Assigned Numbers Authority".) Doing the math results in a reduction of 3,080,192 class C blocks to be removed from the scan list, leaving us with a theoretical list of 13,697,024 blocks.
Let's keep to the facts and look at the numbers. Whichever way you look at it, the Internet is dying. The writing is on the wall: the future for the Internet looks bleak.
Yes but Microsoft software often makes it especially easy for asinine crackheads to do stupid things. For example, running as local administrator by default in some Windows versions, or allowing write access to the Windows system directory even for non-admins. Disguising running an executable as 'opening' a document, so most users cannot tell which attachments are safe to read and which aren't. Executing scripting languages attached to documents so that even viewing a doc can expose you to macro viruses (partly fixed in newer Office versions by prompting about macros, but why is the macro language allowed unrestricted access to modify other documents anyway?). File formats that include leftover binary garbage from old documents, so people can find information you didn't intend to send them.
A Unix-like system with text-based file formats (so you can see exactly what you're sending), a clear separation between root and ordinary user accounts, between editing a document and running a program, and with a traditional text-based mail reader would do a lot more to protect stupid users from themselves. (And Unix is not exactly great in this regard - it's only when you stand it next to Windows that it looks good.)
So it's an unfortunate coincidence that the OS used by the greatest number of idiots is also the one that does the least to stop them doing stupid things. Or perhaps not such a coincidence after all...
And of course the first consumer 3D accelerator is not the first PC 3D accelerator. The IrisVision
dates from 1987 or so. And even that may not be the first.
Remember, nothing new has been invented since X, where X is some surprisingly early date.
Slashdot Mirror
Cool URIs don't change
A bit over-idealistic, but worth aiming towards even if you don't achieve 100% non-URI-breakage in practice.
I feel that search engines should slightly penalize sites that have a history of breaking links or making them redirect to a completely irrelevant page: partly because there is just less chance that the link you follow from the search engine will have the content you want, and partly because even if you do get to a correct page, its usefulness as a bookmark or a link from your own dcuments is reduced.
They make no claim to assess whether music is 'good' or 'crap'. All they try to predict is whether it will sell.
If 'what's hot now' stops selling in large numbers then the algorithm will be adjusted - presumably they keep feeding in the latest songs and their sales volumes.
Yeah I wasn't talking about what Javascript _could_ be used for an an ideal world where all web page authors are non-cracksmoking, but about what it has been used for in practice, which is 99% crud. And that has been the case right back to 1994.
So what's an example of something useful (not advertising or moronic animations) that can be done in Javascript but not done using mid-1990s HTML? Form validation is one example, that would be the one per cent opposed to the 99% of crap I mentioned. But anything apart from that?
I've never understood why sites such as Slashdot want to use cookies for authorization... what's wrong with the standard username/password dialogue box that the browser is able to pop up and which servers like Apache easily support? That means a consistent user interface for logging in across sites, and if the user presses Cancel then he's taken to a sign-up page. The New York Times site seems to be one of the few that does this right.
As far as I recall 99% of Javascript has always been crap and useless and a security hole right from the day it was introduced; the part that has only happened recently is browsers (eg Mozilla Firebird) being sensible enough to block popups and other egregious behaviour. But the reason for that is that during the 1990s there was a too-cosy between the big browser makers (Netscape, Microsoft) and the advertisers that made money (or at least had a good stock price while losing money) through crappy 'portals'.
Ask Igor seems to be an implementation of delta debugging. You can use the delta program to implement this on your own system. You choose a test program (or 'harness') and an input that causes the harness to exit with success; for example your harness might run some executable and test to see if it segfaults - if so success. Then you give an initial input that passes the test (eg causes the segfault). Delta chops out lines of text to find a minimal (or at least 1-minimal, see the website) test case that passes the test (causes the segfault).
This is slightly different to Ask Igor, which takes two different files and finds the important difference between them. But similar in spirit (if much simpler). Apparently the Ask Igor code will be made available for download after it has been used 1000 times from the website.
What exactly causes this phenomenon to be specific to Ipods? It could happen equally well with walkmen or anything having a headphone socket. Is it just because Ipod users are more likely to blog about it and to read others' blogs, causing the practice to take off?
BTW - Ipods are quite expensive, this sounds like a good way to mug people for them.
A separate paper trail is good, but it would seem simpler to just make the main voting record be paper. There are plenty of automated multiple-choice tests where you fill in a box with a pencil and the machine reads your answers. This could be done for voting and realize just the same cost savings as touchscreen voting, but with a built-in paper audit trail. Furthermore, the ballots could be processed by two systems from competing vendors to make sure the results agree. As a last resort they can be counted by hand.
I just don't understand why touch screens are so favoured compared to paper.
So which came first: the Cornell project, or the Fedora project at the University of Hawaii? If Cornell's project was first, why haven't they complained about the name clash until now?
If you are running the software on your own laptop, then you don't need to care about 'vulnerabilities'. It's your laptop so you can run what you like on it anyway! For example, run the exam software in a VMware virtual machine or under a debugger. If you demonstrate this point to the exam organizers I'm sure they will rethink their plan.
To do online exams you need to control the PCs being used, as done with Lexis.
So let me get this straight. NASA has installed a big new Linux-based supercomputer, consuming only 5.2kW of power per rack, that has broken previous transaction processing benchmarks running Oracle 10g.
It should be possible to apply the same technique to other performance-tuning things like buffer sizes, or even choosing between two different algorithms for the same task. Perhaps a program could have a single header file performance.h containing // buffer size for reading // amount of new memory to allocate in a chunk // threshold for switching from quicksort to // merge sort when you have this few elements //
const size_t read_bufsiz = 1000;
const size_t chunk_size = 5000;
const unsigned int merge_sort_thresh = 10;
and then these values would be tuned at the same time as compiler flags.
Sorry, I don't know where ISOs of RHEL are available for download. (That in itself wouldn't stop it being free - for example, the FSF does not provide ISO images of GNU software for download but they will be happy to sell you a physical CD for $5000.)
Doesn't the above apply only if you sign a support contract with Red Hat? Of course you can sign a contract with another company for any stupid conditions you want, but you could also get the software without restrictions and without support (which puts it in the same category as, say, Slackware). If Red Hat won't sell you support except under draconian rules, don't buy it from them.
Yes, Netscape made a scripting language and Sun was happy to stick a meaningless 'Java' label on it although it has nothing to do with Java. Sun agreed to let Netscape use the trademark.
In this case the obvious cure is to render the 'HTML' to plain text first and then do spam-checking on that. Of course if you use a lame mail reader that really wants to display the lovely red colours and FONT SIZE="+9" then you still have a mismatch between what is checked and what is displayed, but not such a big one.
Why not Linux's remote block device? It might perform better than NFS because (assuming read-only access to the block device on the server, with a smaller device for /var) blocks can be cached on the client. (OTOH, a clever NFS implementation could cache bits of filesystem if you promised it the data was not going to change.)
So is there an authentication system for NFS that doesn't trust the client PC? One that is secure against plugging in your own laptop and setting it to have the same IP address and MAC address as an existing machine (which you unplug at the same time)?
I believe that in Sun's current worldview Java is a meaningless trademark to be slapped onto anything: hence Java Desktop. We should have seen the warning signs early with Javascript.
This is Euro-legislation. It's the result of lobbying in Brussels. Then a vote by ministers from all EU countries to adopt the directive - you know, France, Belgium, Germany, all those countries that are far up Bush's arse.
Let's keep to the facts and look at the numbers. Whichever way you look at it, the Internet is dying. The writing is on the wall: the future for the Internet looks bleak.
I have my cvsroot/ directory in my home directory. So I obviously can't put _that_ in CVS...
Yes but Microsoft software often makes it especially easy for asinine crackheads to do stupid things. For example, running as local administrator by default in some Windows versions, or allowing write access to the Windows system directory even for non-admins. Disguising running an executable as 'opening' a document, so most users cannot tell which attachments are safe to read and which aren't. Executing scripting languages attached to documents so that even viewing a doc can expose you to macro viruses (partly fixed in newer Office versions by prompting about macros, but why is the macro language allowed unrestricted access to modify other documents anyway?). File formats that include leftover binary garbage from old documents, so people can find information you didn't intend to send them.
A Unix-like system with text-based file formats (so you can see exactly what you're sending), a clear separation between root and ordinary user accounts, between editing a document and running a program, and with a traditional text-based mail reader would do a lot more to protect stupid users from themselves. (And Unix is not exactly great in this regard - it's only when you stand it next
to Windows that it looks good.)
So it's an unfortunate coincidence that the OS used by the greatest number of idiots is also the one that does the least to stop them doing stupid things. Or perhaps not such a coincidence after all...
And of course the first consumer 3D accelerator is not the first PC 3D accelerator. The IrisVision dates from 1987 or so. And even that may not be the first. Remember, nothing new has been invented since X, where X is some surprisingly early date.