Borges wasn't a linguist, he was a novelist. You know, fiction, the stuff they don't sell at O'Reilly ? The point of the story wasn't to demonstrate a working language anyway, but to raise the idea (and many others) of "languages without nouns"
More geeks should read Borges, he's just what many of us always wanted from Ballard, but somehow never found.
As for Forth (as RPN is less codified as a written language), then "2" certainly isn't a noun. The state of the stack is modified to contain the attribute of "2-ness" as its top value, but nowhere is there a concrete instance of "the ur-2".
Possibly Forth does have a noun, but it only has one of them and it's called "stack".
So it really wouldn't make any difference what language the rules were in
That's a very interesting point, because BNF doesn't really describe the "rules" of a language, but merely the syntax. As the syntax is relatively trivial, then it's entirely possible (as you describe) to replace the "English" serialization with a "Russian" serialization. This is equally easy for C (which uses many punctuation symbols as lexical elements) as it is for Pascal (which uses words). Each "word" is simply a lexical marker for a token.
It's also why C and Pascal programmers can cross-train very easily, but C and Lisp programmers find it much harder. The words are different, but (apart from passing by reference) C and Pascal are very similar. Lisp needs whole new concepts, not just syntactic fiddling.
AFAIK every programming language can be set down, albeit long and complex, in a set of rules in BNF
Not every one, although those that can't (and why) are themselves a fascinating subject (start finding out by reading Godel Escher, Bach)
(Bacchus-Naur Form)
That's just used for pissed-up CS geeks to design new languages on beermats 8-)
There's more to a programming language than the words (or lexical tokens) that describe the syntax. The interesting difference is not whether the reserved word is spelled IF or SI, but the semantic differences of the language's underlying structure. Don't confuse the semantic data model with the mere serialisation (as you chose to mention XML).
Does German's infamous "verb at the end" form encourage functional programming ?
Is the Chinese context-dependent re-use of an pictogram a model of polymorphism ?
Are languages without a first-person inherently supportive of multi-processor systems ?
Could you program in Welsh, just by adding the suffix "io" to all the machine code opcodes ? 8-)
Last night I was reading Jorge Luis Borges' collection of short stories, Labyrinths. One story involves (amongst other things) a language without nouns like "moon", instead using verbs to describe the action of "to moonate" or the state of "mooning".
It struck me whilst reading that this wasn't too far from a stack-oriented language, like RPN or Forth. Unusually this is a case where a concept was probably more familiar to me now than it was to Borges (famously fond of obscure references) when he wrote it.
Don't be an idiot. Mountaineers / skiers who want to be tracked can already choose to wear any number of tracking / alarm devices. These are useful things that don't infringe anyone's privacy.
This device is entirely different. It's implanted, meaning that it's potentially always on (if you want to turn it off, take it off). There is no good reason whatsoever why anyone who wants a temporary usage of this technology (the mountaineer) should need to have it implanted.
The potential for government abuse of this device is enormous. It's not registering myself that I'm worried about (as an economically active, middle-class member of the voting public the politicians won't mess with me too much), but the more disadvantaged of society; the elderly (Help, I've fallen and can't get up) could find that private health sells their privacy in favour of lower cost housing for them in "battery farms for grannies". Petty criminals "tagged" with one of these will find honest work hard to find when the scar is visible (and if you can't get an honest job, what else do you turn to ?)
The US government is also profoundly unworthy of trust. The history of the last century is full of instances of it behaving in a vile manner towards groups of helpless citizens; whether they be the mentally ill, racial groups, Hollywood Communists or just plain poor.
The problem with a "Nanny State" is when "Nanny" turns out to be J Edgar Hoover, wearing a cocktail dress.
Lets also remember the last government that favoured wearing easily readable identification markings....
Your boot drive goes in NTFS, and the ACL is locked down so that some pillock MCSE with a copy of Outhouse can't get whacked by I-LOVE-YOU and take out the system files.
FAT on anything is stupid these days.
I work with an MCSE, who's a loaner from our central IT people (I suspect I know why). A few weeks ago he pulled a floppy from the recycle bucket and stuck it into one of the streaming servers. Fortunately Norton caught the boot sector virus on it, but if he ever pulls a stunt like that again, or if he does it to my SQL box, then I'll cut his fingers off and feed them to badgers.
Having an MCSE doesn't make you an idiot, but there's a large group out there who think they're clueful solely because they have an MCSE. These are dangerous and shouldn't be left with sharp objects, let alone computers.
There's a lot more to the.cx story than is obvious. The people who made the sensible offers some time ago (and the happy smiling people of the charming South Seas island of Bollo) aren't the same registrars as the current money-grabbing leeches.
It's potentially a criminal offence to collect user's CC numbers, post them to alt.warez and bill a bunch of Albanian badger-porn to them. This is unlikely to stop sites collecting CC numbers though.
The DPA is pretty toothless for protecting against privacy issues in today's automated data-capturing environment. It requires some degree of "consent", some requirements on careful storage, and some requirement for the subject's ability to review what is stored. As for defining what's a legitimate business purpose for collecting the data, and what's a gross invasion of privacy, then it's silent. IMHO, we'll never see a general bill of this nature that ever tries to define this issue, unless there's a mechanism (like P3P) that allows the user to negotiate the specifics of privacy with the site, on a per-access basis (and the extent of disclosure permitted thus becomes the subject of a contract).
The DPA 1998 Schedule 2, 2 (b) states one of the conditions for processing to be "necessary"
(b) for the taking of steps at the request of the data subject with a view to entering into a contract.
Any contract-drafting bottom-feeding lawshark can present a retail site such that it's accessed "with a view to entering into a contract".
I use different machines - there's no other way. IE versions won't co-exist (they claim it now works for 4 & 5.5, but I found it very unstable). You also need Macs and Linux to run the non-Windows browsers. Even the "same" browser on a different OS can show different rendering bugs; especially support of CSS features and font sizing.
The Bookmarklets site has some cute tools for testing at different window sizes, with the minimum of effort.
The deja-browser site gives a feel for the "experience" of old browsers, but it doesn't emulate the bugs accurately enough to be used for compatibility testing.
Compatibility with "old" browsers is easy anyway; anything looks flat, grey and ugly through Mosaic 1.0, but then it just deosn't get any better. The real hassles are finding out the little things like IE4 not supporting "float" in CSS.
I'm not a pro, [...] conforms with the standards, [...] If the browsers can't handle that, *$%^#* them.
I am a pro, and sadly I can't get away with this. My pages fall into two groups; those where I need to make them work cross-browser, and those(generally the personal stuff) where I can take the same sensible line you do (pure XHTML, on any of the DTDs, and only CSS for formatting).
Allow between 5 and 10 times more effort to develop a decently cross-browser page, rather than nice simple CSS.
company is in fact basing their entire online strategy on Javascript.
This is a good strategy (especially with client-side XML and DXML), provided that your site is either an intranet, or sufficiently valuable to the users to make them permit scripting on it. If, for instance, I were a financial institution (and they're already trusting me with their portfolio) then I'd have no problem at all in going all out for Javascript.
As for security, I'm _much_ less worried about script than I am about ActiveXs. It's very difficult, and reliant on obscure holes, to do damage with Javascript. OTOH, any monkey with a copy of VB can make an unsigned ActiveX that trashes the whole box (it's evil, and it's obvious, but it's still do-able).
Javascript is also entirely standard, for the simple reason that there's only one browser left out there; IE5.5 - Market share for non-IE is damn near negligible, and IE 4 or 5 users will easily be pushed to upgrade (IE5 especially - 5.5 is much more stable).
I went through a dose of this a while ago. I had a really interesting project, but I was just sat at home staring at the screen and doing anything from housework to accountancy to avoid getting down to the code.
I was just sick of working on my own for too many years.
What did I do ? Went back out on the contract circuit, but this time I was lucky enough to have a slot at an interesting site, not just another contractor sausage-factory. Right now I'm working for HP Labs; an enormous shed filled with the smartest bunch of people I've ever met in one place. Pick a cube, any cube, and there will be someone in there doing interesting stuff, with interesting ideas to share. Sitting on your own all day gets to you in the end, even with net.access.
After this, I'm fired with enthusiam to go and do anything -- except I don't want to leave 8-)
How do you fix you immediate problem ? Read some other poster's ideas, but basically you need to get it fixed and get that project out of the way (painful though it is). For the next project though, I'd look seriously at finding some work with a new bunch of interesting people. Why not work in Europe, on a 6 month fixed-term deal, just for a change of scenery ?
The Geek Lifestyle is fun while it lasts, but there comes a day when you want more out of life than an excuse for unlimited O'Reillys and Pizza.
Looks almost like they're beating swords into plowshares
(a hollow voice says "plough") 8-)
More like taking money from the currently-rich ploughshare venture capitalists and using it to fund the next cycle's sword research.
The real problem with AngelHalo is that it's a glorified cellphone system for coverage over the USA's enormous road network. It's a great system, where it covers, but it's targeted at where the traffic is, not a truly global system.
Iridium works everywhere, even when the local technology doesn't extend beyond the cow-dung fire and the AK47. I only know a few Iridium users, all aid workers in the absolute arse-ends of the world (Northern Afghanistan right this minute). They'd have a lot of trouble without their Iridia (sp?), as there just isn't an alternative, other than the military's comsat networks.
What made IE so Succesfull and capture the mass market? Microsoft's huge marketing machine.
Nothing made IE a success, except Netscape making itself a failure.
We didn't switch to IE because it was good, we moved because Netscape was worse and there wasn't any other choice. Only long afterwards did site authors start to use the DHTML and other bells and whistles that IE brought with it.
*sigh* What's that saying about a little bit of knowledge being a dangerous thing?
Absolutely. Which is why John's comment is so relevant -- Let's suppose you're working for a Humorless Corporate, and you're found running Ethereal. No admin is going to be happy about it, most will go ballistic, and any pointy-hair is going to see it as a major end-of-job crime. In almost every site I've ever worked, that's a hanging offence for certain (in one Scottish waterfront site, I think it's still one of the few literally hanging offences on the British statute book !).
Are these people right ? How do I know; it all depends on what you did with it, but nearly every company, nearly everywhere, is going to see this as distinctly A Bad Thing.
I'm considering running it at work -- I'm working on streaming, and sniffing traffic from my servers would be pretty useful. Out of courtesy I'd warn the admins though, and I'm lucky in that ours are clueful enough to understand why this is reasonable.
What's the "platform" ? To me, it's the combination of browser and OS, not just having the same menu layout. With my web-author hat on, I care about platform issues like "Do my users have web browsers with reasonable client-side XSL ?". I don't give a damn about the user interface; that's the user's problem and it won't break my pages.
If Mozilla's "platform" isn't as capable as IE's, I'm not going to start using it. If it's no better than IE's, or it doesn't have a huge market share, then I'm not going to write for it.
You don't need HTML editing to fill out a webform. And I think we can ALL agree that HTML mail is a bug, not a feature
Amen to that.
How did HTML editing get in there anyway ? If it's a "cool feature" that some over-zealous geek thought should be in there, despite the deadline pressures, then why HTML editing ? Is there a single web-monkey out there, worthy of their prehensile tail, that needs or uses a HTML editor more complex than a basic text editor ? Scrape 'n Dump Tables ? That's a freebie for Runtpage and "My Little HomePage" on Geocities, not a tool that a real HTML geek needs, wants or would use.
Besides which, HTML authoring is dead -- XHTML is where it's at, and the faster we can start ignoring the legacy browsers, the easier XHTML/CSS authoring gets. If you aren't already doing it, start writing compliant XHTML today (no, it isn't incompatible)
We should count our blessing that some little pointy-beard didn't think that a Flash editor would be cool too.
Mozilla allows Cascading Style sheet to format XML (and soon, XSLT). This is more than cool. This is the future.
Bad news. XML+CSS has been dead for at least a year. XML is the future of browsers, but it's going to be XSLT that drives it, not CSS. CSS is a pitiful means of delivering XML, even if the CSS rendering engine does work right. There are loads of useful, real-world, right-this-minute, client-side XML uses that CSS can't even begin to address. Mozilla has missed the boat on this one.
this discussion is whether to submit bug reports to the appropriate authorities in a manner that will get holes plugged and fixes distributed to admins
I'm all in favour of controlled distribution of exploits, so that there's a chance to fix them before they're commonplace and (mis-)used.
However this hasn't worked. Outlook is still wide-open (and anti-virus sniffers are not the right solution to this). Routers are still misconfigured. The sad fact is that there aren't enough admins about who even know who CERT is, and these dummies need a wake-up call. We've tried the quiet route of patching holes before they're widespread, and it just isn't working. Too many machines on the modern Net simply stops "the old ways" from scaling up.
Last week, someone (reputedly an "admin", although I use the term with caution) stuck a floppy with a boot sector virus into one of my major servers. They still think that, because the scanner saw it before they caused an infection, that it's OK to re-use unknown floppies from the wastebins and use them to transfer data onto a production server. LART, LART, LART !
Slashdot Mirror
Borges wasn't a linguist, he was a novelist. You know, fiction , the stuff they don't sell at O'Reilly ? The point of the story wasn't to demonstrate a working language anyway, but to raise the idea (and many others) of "languages without nouns"
More geeks should read Borges, he's just what many of us always wanted from Ballard, but somehow never found.
As for Forth (as RPN is less codified as a written language), then "2" certainly isn't a noun. The state of the stack is modified to contain the attribute of "2-ness" as its top value, but nowhere is there a concrete instance of "the ur-2".
Possibly Forth does have a noun, but it only has one of them and it's called "stack".
It's trivially easy to find new primes.
What's hard is to take an arbitrary large number, then see if it is prime or not.
So it really wouldn't make any difference what language the rules were in
That's a very interesting point, because BNF doesn't really describe the "rules" of a language, but merely the syntax. As the syntax is relatively trivial, then it's entirely possible (as you describe) to replace the "English" serialization with a "Russian" serialization. This is equally easy for C (which uses many punctuation symbols as lexical elements) as it is for Pascal (which uses words). Each "word" is simply a lexical marker for a token.
It's also why C and Pascal programmers can cross-train very easily, but C and Lisp programmers find it much harder. The words are different, but (apart from passing by reference) C and Pascal are very similar. Lisp needs whole new concepts, not just syntactic fiddling.
AFAIK every programming language can be set down, albeit long and complex, in a set of rules in BNF
Not every one, although those that can't (and why) are themselves a fascinating subject (start finding out by reading Godel Escher, Bach)
(Bacchus-Naur Form)
That's just used for pissed-up CS geeks to design new languages on beermats 8-)
Real Programmers can write FORTRAN in any language.
There's more to a programming language than the words (or lexical tokens) that describe the syntax. The interesting difference is not whether the reserved word is spelled IF or SI, but the semantic differences of the language's underlying structure. Don't confuse the semantic data model with the mere serialisation (as you chose to mention XML).
Does German's infamous "verb at the end" form encourage functional programming ?
Is the Chinese context-dependent re-use of an pictogram a model of polymorphism ?
Are languages without a first-person inherently supportive of multi-processor systems ?
Could you program in Welsh, just by adding the suffix "io" to all the machine code opcodes ? 8-)
Last night I was reading Jorge Luis Borges' collection of short stories, Labyrinths. One story involves (amongst other things) a language without nouns like "moon", instead using verbs to describe the action of "to moonate" or the state of "mooning".
It struck me whilst reading that this wasn't too far from a stack-oriented language, like RPN or Forth. Unusually this is a case where a concept was probably more familiar to me now than it was to Borges (famously fond of obscure references) when he wrote it.
Mozilla is more bloated than IE, slower, and less credible to the suits than even Netscape.
We need a stable web browser
Don't be an idiot. Mountaineers / skiers who want to be tracked can already choose to wear any number of tracking / alarm devices. These are useful things that don't infringe anyone's privacy.
This device is entirely different. It's implanted, meaning that it's potentially always on (if you want to turn it off, take it off). There is no good reason whatsoever why anyone who wants a temporary usage of this technology (the mountaineer) should need to have it implanted.
The potential for government abuse of this device is enormous. It's not registering myself that I'm worried about (as an economically active, middle-class member of the voting public the politicians won't mess with me too much), but the more disadvantaged of society; the elderly (Help, I've fallen and can't get up) could find that private health sells their privacy in favour of lower cost housing for them in "battery farms for grannies". Petty criminals "tagged" with one of these will find honest work hard to find when the scar is visible (and if you can't get an honest job, what else do you turn to ?)
The US government is also profoundly unworthy of trust. The history of the last century is full of instances of it behaving in a vile manner towards groups of helpless citizens; whether they be the mentally ill, racial groups, Hollywood Communists or just plain poor.
The problem with a "Nanny State" is when "Nanny" turns out to be J Edgar Hoover, wearing a cocktail dress.
Lets also remember the last government that favoured wearing easily readable identification markings....
My Cat is more wired than Kevin Warwick.
At least he can automatically open catflaps.
Your boot drive goes in NTFS, and the ACL is locked down so that some pillock MCSE with a copy of Outhouse can't get whacked by I-LOVE-YOU and take out the system files.
FAT on anything is stupid these days.
I work with an MCSE, who's a loaner from our central IT people (I suspect I know why). A few weeks ago he pulled a floppy from the recycle bucket and stuck it into one of the streaming servers. Fortunately Norton caught the boot sector virus on it, but if he ever pulls a stunt like that again, or if he does it to my SQL box, then I'll cut his fingers off and feed them to badgers.
Having an MCSE doesn't make you an idiot, but there's a large group out there who think they're clueful solely because they have an MCSE. These are dangerous and shouldn't be left with sharp objects, let alone computers.
There's a lot more to the .cx story than is obvious. The people who made the sensible offers some time ago (and the happy smiling people of the charming South Seas island of Bollo) aren't the same registrars as the current money-grabbing leeches.
I love IE, I'm just not so keen on the Borg
Mozilla sucks mightily (IMHO, lets not get sidetracked), but it's the only competition that IE has, something we all desperately need.
In a world without any IE competition, IE will soon look like nothing more than a WebTV plugging us straight into Shop@Redmond and The Barney Channel.
It's potentially a criminal offence to collect user's CC numbers, post them to alt.warez and bill a bunch of Albanian badger-porn to them. This is unlikely to stop sites collecting CC numbers though.
The DPA is pretty toothless for protecting against privacy issues in today's automated data-capturing environment. It requires some degree of "consent", some requirements on careful storage, and some requirement for the subject's ability to review what is stored. As for defining what's a legitimate business purpose for collecting the data, and what's a gross invasion of privacy, then it's silent. IMHO, we'll never see a general bill of this nature that ever tries to define this issue, unless there's a mechanism (like P3P) that allows the user to negotiate the specifics of privacy with the site, on a per-access basis (and the extent of disclosure permitted thus becomes the subject of a contract).
The DPA 1998 Schedule 2, 2 (b) states one of the conditions for processing to be "necessary"
(b) for the taking of steps at the request of the data subject with a view to entering into a contract.
Any contract-drafting bottom-feeding lawshark can present a retail site such that it's accessed "with a view to entering into a contract".
without using 4 different PCs
I use different machines - there's no other way. IE versions won't co-exist (they claim it now works for 4 & 5.5, but I found it very unstable). You also need Macs and Linux to run the non-Windows browsers. Even the "same" browser on a different OS can show different rendering bugs; especially support of CSS features and font sizing.
The Bookmarklets site has some cute tools for testing at different window sizes, with the minimum of effort.
The deja-browser site gives a feel for the "experience" of old browsers, but it doesn't emulate the bugs accurately enough to be used for compatibility testing.
Compatibility with "old" browsers is easy anyway; anything looks flat, grey and ugly through Mosaic 1.0, but then it just deosn't get any better. The real hassles are finding out the little things like IE4 not supporting "float" in CSS.
I'm not a pro, [...] conforms with the standards, [...] If the browsers can't handle that, *$%^#* them.
I am a pro, and sadly I can't get away with this. My pages fall into two groups; those where I need to make them work cross-browser, and those(generally the personal stuff) where I can take the same sensible line you do (pure XHTML, on any of the DTDs, and only CSS for formatting).
Allow between 5 and 10 times more effort to develop a decently cross-browser page, rather than nice simple CSS.
company is in fact basing their entire online strategy on Javascript.
This is a good strategy (especially with client-side XML and DXML), provided that your site is either an intranet, or sufficiently valuable to the users to make them permit scripting on it. If, for instance, I were a financial institution (and they're already trusting me with their portfolio) then I'd have no problem at all in going all out for Javascript.
As for security, I'm _much_ less worried about script than I am about ActiveXs. It's very difficult, and reliant on obscure holes, to do damage with Javascript. OTOH, any monkey with a copy of VB can make an unsigned ActiveX that trashes the whole box (it's evil, and it's obvious, but it's still do-able).
Javascript is also entirely standard, for the simple reason that there's only one browser left out there; IE5.5 - Market share for non-IE is damn near negligible, and IE 4 or 5 users will easily be pushed to upgrade (IE5 especially - 5.5 is much more stable).
I went through a dose of this a while ago. I had a really interesting project, but I was just sat at home staring at the screen and doing anything from housework to accountancy to avoid getting down to the code.
I was just sick of working on my own for too many years.
What did I do ? Went back out on the contract circuit, but this time I was lucky enough to have a slot at an interesting site, not just another contractor sausage-factory. Right now I'm working for HP Labs; an enormous shed filled with the smartest bunch of people I've ever met in one place. Pick a cube, any cube, and there will be someone in there doing interesting stuff, with interesting ideas to share. Sitting on your own all day gets to you in the end, even with net.access.
After this, I'm fired with enthusiam to go and do anything -- except I don't want to leave 8-)
How do you fix you immediate problem ? Read some other poster's ideas, but basically you need to get it fixed and get that project out of the way (painful though it is). For the next project though, I'd look seriously at finding some work with a new bunch of interesting people. Why not work in Europe, on a 6 month fixed-term deal, just for a change of scenery ?
The Geek Lifestyle is fun while it lasts, but there comes a day when you want more out of life than an excuse for unlimited O'Reillys and Pizza.
Looks almost like they're beating swords into plowshares
(a hollow voice says "plough") 8-)
More like taking money from the currently-rich ploughshare venture capitalists and using it to fund the next cycle's sword research.
The real problem with AngelHalo is that it's a glorified cellphone system for coverage over the USA's enormous road network. It's a great system, where it covers, but it's targeted at where the traffic is, not a truly global system.
Iridium works everywhere, even when the local technology doesn't extend beyond the cow-dung fire and the AK47. I only know a few Iridium users, all aid workers in the absolute arse-ends of the world (Northern Afghanistan right this minute). They'd have a lot of trouble without their Iridia (sp?), as there just isn't an alternative, other than the military's comsat networks.
What made IE so Succesfull and capture the mass market? Microsoft's huge marketing machine.
Nothing made IE a success, except Netscape making itself a failure.
We didn't switch to IE because it was good, we moved because Netscape was worse and there wasn't any other choice. Only long afterwards did site authors start to use the DHTML and other bells and whistles that IE brought with it.
*sigh* What's that saying about a little bit of knowledge being a dangerous thing?
Absolutely. Which is why John's comment is so relevant -- Let's suppose you're working for a Humorless Corporate, and you're found running Ethereal. No admin is going to be happy about it, most will go ballistic, and any pointy-hair is going to see it as a major end-of-job crime. In almost every site I've ever worked, that's a hanging offence for certain (in one Scottish waterfront site, I think it's still one of the few literally hanging offences on the British statute book !).
Are these people right ? How do I know; it all depends on what you did with it, but nearly every company, nearly everywhere, is going to see this as distinctly A Bad Thing.
I'm considering running it at work -- I'm working on streaming, and sniffing traffic from my servers would be pretty useful. Out of courtesy I'd warn the admins though, and I'm lucky in that ours are clueful enough to understand why this is reasonable.
Mozilla/XUL really is cross platform.
What's the "platform" ? To me, it's the combination of browser and OS, not just having the same menu layout. With my web-author hat on, I care about platform issues like "Do my users have web browsers with reasonable client-side XSL ?". I don't give a damn about the user interface; that's the user's problem and it won't break my pages.
If Mozilla's "platform" isn't as capable as IE's, I'm not going to start using it. If it's no better than IE's, or it doesn't have a huge market share, then I'm not going to write for it.
You don't need HTML editing to fill out a webform. And I think we can ALL agree that HTML mail is a bug, not a feature
Amen to that.
How did HTML editing get in there anyway ? If it's a "cool feature" that some over-zealous geek thought should be in there, despite the deadline pressures, then why HTML editing ? Is there a single web-monkey out there, worthy of their prehensile tail, that needs or uses a HTML editor more complex than a basic text editor ? Scrape 'n Dump Tables ? That's a freebie for Runtpage and "My Little HomePage" on Geocities, not a tool that a real HTML geek needs, wants or would use.
Besides which, HTML authoring is dead -- XHTML is where it's at, and the faster we can start ignoring the legacy browsers, the easier XHTML/CSS authoring gets. If you aren't already doing it, start writing compliant XHTML today (no, it isn't incompatible)
We should count our blessing that some little pointy-beard didn't think that a Flash editor would be cool too.
Opera doesn't do Unicode. Last year that was bad news, this year (when XML is everywhere) it's terminal.
Mozilla allows Cascading Style sheet to format XML (and soon, XSLT). This is more than cool. This is the future.
Bad news. XML+CSS has been dead for at least a year. XML is the future of browsers, but it's going to be XSLT that drives it, not CSS. CSS is a pitiful means of delivering XML, even if the CSS rendering engine does work right. There are loads of useful, real-world, right-this-minute, client-side XML uses that CSS can't even begin to address. Mozilla has missed the boat on this one.
this discussion is whether to submit bug reports to the appropriate authorities in a manner that will get holes plugged and fixes distributed to admins
I'm all in favour of controlled distribution of exploits, so that there's a chance to fix them before they're commonplace and (mis-)used.
However this hasn't worked. Outlook is still wide-open (and anti-virus sniffers are not the right solution to this). Routers are still misconfigured. The sad fact is that there aren't enough admins about who even know who CERT is, and these dummies need a wake-up call. We've tried the quiet route of patching holes before they're widespread, and it just isn't working. Too many machines on the modern Net simply stops "the old ways" from scaling up.
Last week, someone (reputedly an "admin", although I use the term with caution) stuck a floppy with a boot sector virus into one of my major servers. They still think that, because the scanner saw it before they caused an infection, that it's OK to re-use unknown floppies from the wastebins and use them to transfer data onto a production server. LART, LART, LART !