This is the technique used by portsnap; basically you generate binary diffs from a known starting point, and the client keeps track of what new patches it needs to keep in sync. Since you're just serving static files, scaling it should be as easy and cheap as it gets.
rsync is highly general purpose; your servers will end up generating hashes for every n-bytes of every file for every client, which is a lot more heavyweight than just serving patches you generate once. SubVersion may be more effecient since it should know something about the files it's checked out previously, but it's still going to end up dynamically generating diffs between whatever versions each client has and the latest; this likely gets worse if your clients aren't tracking HEAD.
Also note that a custom solution can likely get away with a single tag file detailing the latest patches; rsync and svn are going to be scanning their directory trees religiously. Both you and your users will probably appreciate a single GET to a small file on a webserver than a load of CPU use and disk thrashing.
Yup, that's kinda why I said "I wonder why it hasn't taken off", given so many people were really itching for the new features (even if they removed half of them during ZE2 development, meh). Oh, and you didn't explicitly say anything about usability, just "cleaner";)
Seems a bit like a chicken and egg problem in some ways; very few people use the new features because most people don't have control over their own servers, so those that do have control don't see a point in upgrading because.. nobody uses the new features. Having a few big-name apps start requiring them and making obvious the benefits would probably go a long way towards helping the demand side.
Not that I don't see it as a good thing to see users find better languages for them; a monoculture where "web scripting" == PHP really isn't healthy, I don't care how free it is. Competition makes everything suck less:)
"The error handling code works and I challenge you to find a cleaner way to let the developer know exactly where an error occured so they can fix it."
I'll take "exceptions" for £50, Bob. Unfortunate that PHP5 doesn't seem to be taking off quite like PHP4 did; I wonder if that's because many of the people who would find the new new features attractive are finding other languages suit them better? I know I did, and I used to really love PHP and was dead excited by Zend Engine 2:/
One of my systems has 4 drives mounted on 3.5" - 5.25" brackets, with a single 120mm fan in front of them; keeps them around 35c, which is a good deal better than the 50c badly ventilated drives seem to live at.
The better solution is of course a case that doesn't suck and let your drives melt, but finding them can be tricky, especially when you want them to be fairly cheap and not suck in other ways.
You do? What radioactive byproducts do we produce that's both dangerously radioactive and remains that way for such a long time? Or did you just pick a random large number that sounded good?
Hm? I'm not free to consume cannabis, otherwise I'd be able to buy it from the corner shop next to the alcohol and tobacco, or grow it in my cupboard or get a bit for my friends without risking having my life ruined. I'd be able to relax with a joint in the park on a nice day without risking a warning and having it stolen from me by the very people who are supposed to protect me.
Of course I'm slightly more free in this regard than I was a few years ago, so yeah, different flavours. I'm just bitter because I can't find a dealer;)
"What I meant was that IE renders XHTML just fine. And it does."
Mostly. Can you even put in an XML prologue without it screaming and running away to quirks mode yet?:/
"And it isn't good enough reason for folks who do know what they're doing, like you supposedly do, to not use XHTML full stop."
Sure, I'm not saying it is; thing is most XHTML users are in the camp that doesn't really know what they're doing, and even those that do are often constrained by legacy systems and other developers/users which are resistant to being retaught. But seriously, even when that isn't an issue, is there really a big reason to use XHTML 1 because of some vague notion that it'll make migrating to a new spec a few years hence easier? Why not just make sure your web-apps are written nicely enough so it's easier to add closer to when it'll actually be useful, and use that flexibility now for more immediately useful stuff like Atom/RSS feeds? Why introduce unnecessary complexity?
"Actually, you're right. This is pointless."
Well, I dunno, some might say avoiding work is a perfectly valid point;)
"Slashdot should just serve up application/xhtml+xml and lock the IE weenies out"
Unfortunately Slashcode is one of those legacy applications which are resistant to change. Messy Perl and a million user-input HTML elements is hardly the perfect system to be serializing XML from; serving XHTML as XML is all fun and games until something trivial breaks and a million users get a Syntax Error. Be permissive in what you accept indeed...
BTW, is your URL supposed to redirect to some eOpinions TFT review?:)
10's a bit over the top, but I do find I use 7 of the 10 on my MX510 regularly; the obvious 5 from LMB/MMB/RMB/Mouse Wheel Up/Dn, and two thumb buttons for back/forward. The other three I barely even notice.
So, any news on a mouse that doesn't build up nasty greese stains on the buttons?
IE only "supports" XHTML because it, like most other browsers, uses a tag-soup parser which even breaks standards HTML was originally based on. This isn't a good starting point for something that's supposed to be cleaner, especially when you get into the more complex issues associated with dual-mode XHTML-as-HTML/XHTML-as-XML. Makes it easier? Guffaw. All these sites which think they're using XHTML are in for a surprise when they try switching to XML mode and find half their JavaScript breaks, and that really, their CMS isn't that great at serializing well-formed XML. Sites which manage to do it properly may well get a cookie, but "easier" isn't something I think they're going to be touting.
But don't mind me, I've only been doing this for 9 years, what do I know? Or Hixie for that matter, who the hell does that guy think he is?
Thanks for playing Pointless Interwub Argument 4.5. Your turn.
XHTML introduces all sorts of compatibility problems, many of which probably make most purists somewhat nervous. What it adds isn't very (at all?) useful in the vast majority of cases, even ignoring the levels of browser support. Just because it's newer doesn't mean it's better, or that you should use it over an older standard, especially when the newer one is little more than a reformulation of the old one in a new, mostly identical but subtly incompatible language.
This isn't to say XHTML is bad, and that you shouldn't use it, but doing so just because it's newer and gives you fuzzy feelings inside isn't a great reason for favouring it.
Ah, servers, yes. Immortal power supplies and HVAC doesn't come cheap; when you start having to live within a power and BTU budget, being able to drop a few KW per rack is *really* appealing.
This comes hand-in-hand with multiple-cores per chip; server applications tend to have a lot of concurrency, so trading off some raw performance per core for the ability to fit 2-4 on a single chip with a sensible power envelope is a great win. So what if a single core has 30% less throughput if you have twice as many of them to spread load across?
Even for the average user, the reduction in latency and greater ability to multitask probably outweighs a few FPS in Doom 3.
My fan "repair" technique: remove backing (normally a sticker), spray with WD-40 (optional; just to clean it out if necessary), then put in a drop of oil, , make sure it's getting into the bearings, soak up excess, reattach sticker or seal with tape; done.
If a fan seems to be struggling it's a good idea to re-lubricate it like this before the bearings actually get damaged. I always find myself doing this on crappy chipset fans; it can easily double their lifespan.
VFS = also the kernel-layer filesystem API, on which everything from procfs to ext2 is interfaced. Bypassing this layer for certain operations on certain filesystems certainly sounds like something that should be tricky to get merged into mainline.
"I can't stand the rendering implementation in Opera, it's borked (yes seriously it has a lot more bugs than firefox"
This isn't my experience. Obscure clipping bugs, rendering problems and general clunkiness seem to be mainly the realm of IE and Gecko when I'm knocking out CSS and HTML. YMMV, a few anecdotes doesn't exactly make good evidence:)
Ironically when I look up half these bugs in Gecko, the QA contact is a guy who works for Opera QA..:)
"I'm not sure Opera even supports XSLT yet does it?"
No. Am I expected to believe this is a significant omission next to all the other stuff it supports? I'm sure if more than 3 demo websites try using it Opera will adapt. Does Firefox support SVG out of the box yet? Aural stylesheets and voice I/O? BitTorrent? I dare say these are more interesting to more people than an awful XML transformation language most people like purely server-side.
"(something that even IE has supported since version 5!)"
Right. Wasn't it based on working draft specifications and thus pretty much entirely broken by release? I remember it barely being able to parse well formed XML, never mind any languages using it.
I'm in the UK; even the cheapo £30 players Amazon do are generally multiregion, and disabling the region protection on supposedly fixed players seems generally a case of typing in a few sequences on the remote. Do you not even get that over there? Do your TV's at least generally support PAL?
"You've probably heard the SR-71 is a severe leaker, and I'll try to put this into perspective. Once LN2 is serviced a few hours prior to launch, the fuel system becomes pressurized, and that's when the real leaks start. Normally, about five or six steady fuel leaks (about the width of a drinking straw) show up coming from both inboard wings, falling about six feet to the ground. The entire bottom of the fuselage becomes wet, and starts dripping onto the hanger floor. Some puddling starts to accumulate on top of the inboard wings, and at times runs off the wing onto the floor. In some bad leakers, fountains can be seen spraying upward from the top of the inboard wings, ranging anywhere from two inches to three feet in height."
Guy who maintained both planes > Anonymous Coward on/.
Re:I'm sure 99 % /.'ers have better stories...
on
10 Computer Mishaps
·
· Score: 1
Of course these days you could have used/rescue/sh, etc. Should really have been there since day 1...
If your OS lacks such a thing, making one using busybox or so comes highly recommended, remembering to make it statically linked and immutable afterwards:)
Sure, but you'll need to make sure you have a government approved electrical pulse generator implanted into your hippocampus and motor cortex to prevent you forming long term memories or making any copies while doing so.
Where's the lack of control? I don't like Adaptec's attitude with regard to aaccli, but that's their choice; I have the choice not to buy any more of their shitty RAID controllers. Ditto with WiFi cards; I have control over whether I'm happy to have my hardware work even though the software may be closed because those that make it are forced to make it so by dicks, or I can get something good.
Delivering "ultimatums" which everyone knows are going to be ignored just alienates people, users, developers and companies alike. Here in the real world, I need aac(4) to work because they're in a couple of important machines, and I wouldn't appreciate FreeBSD removing it just because the only online management tool is only available from Dell/Red Hat as a Linux binary. What does removing it gain anyone, aside from making Adaptec less likely to fund/support future FreeBSD driver development and pissing off those who actually use the cards?
My appologies for being so pragmatic; however, my attitude is more along the lines of "if company x wants to be lame and only provide binaries, fine, I'll just buy from company y who do provide open source drivers". Part of control is being able to do non-optimal things.
Slashdot Mirror
This is the technique used by portsnap; basically you generate binary diffs from a known starting point, and the client keeps track of what new patches it needs to keep in sync. Since you're just serving static files, scaling it should be as easy and cheap as it gets.
rsync is highly general purpose; your servers will end up generating hashes for every n-bytes of every file for every client, which is a lot more heavyweight than just serving patches you generate once. SubVersion may be more effecient since it should know something about the files it's checked out previously, but it's still going to end up dynamically generating diffs between whatever versions each client has and the latest; this likely gets worse if your clients aren't tracking HEAD.
Also note that a custom solution can likely get away with a single tag file detailing the latest patches; rsync and svn are going to be scanning their directory trees religiously. Both you and your users will probably appreciate a single GET to a small file on a webserver than a load of CPU use and disk thrashing.
Yup, that's kinda why I said "I wonder why it hasn't taken off", given so many people were really itching for the new features (even if they removed half of them during ZE2 development, meh). Oh, and you didn't explicitly say anything about usability, just "cleaner" ;)
:)
Seems a bit like a chicken and egg problem in some ways; very few people use the new features because most people don't have control over their own servers, so those that do have control don't see a point in upgrading because.. nobody uses the new features. Having a few big-name apps start requiring them and making obvious the benefits would probably go a long way towards helping the demand side.
Not that I don't see it as a good thing to see users find better languages for them; a monoculture where "web scripting" == PHP really isn't healthy, I don't care how free it is. Competition makes everything suck less
"The error handling code works and I challenge you to find a cleaner way to let the developer know exactly where an error occured so they can fix it."
:/
I'll take "exceptions" for £50, Bob. Unfortunate that PHP5 doesn't seem to be taking off quite like PHP4 did; I wonder if that's because many of the people who would find the new new features attractive are finding other languages suit them better? I know I did, and I used to really love PHP and was dead excited by Zend Engine 2
"What happens to your data if someone deletes it, and your "backup system" is a RAID array ?"
I restore it from my collection of incremental dumps from said backup RAID array to the live array.
One of my systems has 4 drives mounted on 3.5" - 5.25" brackets, with a single 120mm fan in front of them; keeps them around 35c, which is a good deal better than the 50c badly ventilated drives seem to live at.
The better solution is of course a case that doesn't suck and let your drives melt, but finding them can be tricky, especially when you want them to be fairly cheap and not suck in other ways.
"You have to store it for a billion years"
You do? What radioactive byproducts do we produce that's both dangerously radioactive and remains that way for such a long time? Or did you just pick a random large number that sounded good?
"The freedom to consume soft drugs"
;)
Hm? I'm not free to consume cannabis, otherwise I'd be able to buy it from the corner shop next to the alcohol and tobacco, or grow it in my cupboard or get a bit for my friends without risking having my life ruined. I'd be able to relax with a joint in the park on a nice day without risking a warning and having it stolen from me by the very people who are supposed to protect me.
Of course I'm slightly more free in this regard than I was a few years ago, so yeah, different flavours. I'm just bitter because I can't find a dealer
"What I meant was that IE renders XHTML just fine. And it does."
:/
;)
:)
Mostly. Can you even put in an XML prologue without it screaming and running away to quirks mode yet?
"And it isn't good enough reason for folks who do know what they're doing, like you supposedly do, to not use XHTML full stop."
Sure, I'm not saying it is; thing is most XHTML users are in the camp that doesn't really know what they're doing, and even those that do are often constrained by legacy systems and other developers/users which are resistant to being retaught. But seriously, even when that isn't an issue, is there really a big reason to use XHTML 1 because of some vague notion that it'll make migrating to a new spec a few years hence easier? Why not just make sure your web-apps are written nicely enough so it's easier to add closer to when it'll actually be useful, and use that flexibility now for more immediately useful stuff like Atom/RSS feeds? Why introduce unnecessary complexity?
"Actually, you're right. This is pointless."
Well, I dunno, some might say avoiding work is a perfectly valid point
"Slashdot should just serve up application/xhtml+xml and lock the IE weenies out"
Unfortunately Slashcode is one of those legacy applications which are resistant to change. Messy Perl and a million user-input HTML elements is hardly the perfect system to be serializing XML from; serving XHTML as XML is all fun and games until something trivial breaks and a million users get a Syntax Error. Be permissive in what you accept indeed...
BTW, is your URL supposed to redirect to some eOpinions TFT review?
10's a bit over the top, but I do find I use 7 of the 10 on my MX510 regularly; the obvious 5 from LMB/MMB/RMB/Mouse Wheel Up/Dn, and two thumb buttons for back/forward. The other three I barely even notice.
So, any news on a mouse that doesn't build up nasty greese stains on the buttons?
IE only "supports" XHTML because it, like most other browsers, uses a tag-soup parser which even breaks standards HTML was originally based on. This isn't a good starting point for something that's supposed to be cleaner, especially when you get into the more complex issues associated with dual-mode XHTML-as-HTML/XHTML-as-XML. Makes it easier? Guffaw. All these sites which think they're using XHTML are in for a surprise when they try switching to XML mode and find half their JavaScript breaks, and that really, their CMS isn't that great at serializing well-formed XML. Sites which manage to do it properly may well get a cookie, but "easier" isn't something I think they're going to be touting.
But don't mind me, I've only been doing this for 9 years, what do I know? Or Hixie for that matter, who the hell does that guy think he is?
Thanks for playing Pointless Interwub Argument 4.5. Your turn.
Article looks perfectly reasonable to me.
XHTML introduces all sorts of compatibility problems, many of which probably make most purists somewhat nervous. What it adds isn't very (at all?) useful in the vast majority of cases, even ignoring the levels of browser support. Just because it's newer doesn't mean it's better, or that you should use it over an older standard, especially when the newer one is little more than a reformulation of the old one in a new, mostly identical but subtly incompatible language.
This isn't to say XHTML is bad, and that you shouldn't use it, but doing so just because it's newer and gives you fuzzy feelings inside isn't a great reason for favouring it.
Hixie said it best. Using HTML 4.01 Strict probably makes most purists *happier*.
Ah, servers, yes. Immortal power supplies and HVAC doesn't come cheap; when you start having to live within a power and BTU budget, being able to drop a few KW per rack is *really* appealing.
This comes hand-in-hand with multiple-cores per chip; server applications tend to have a lot of concurrency, so trading off some raw performance per core for the ability to fit 2-4 on a single chip with a sensible power envelope is a great win. So what if a single core has 30% less throughput if you have twice as many of them to spread load across?
Even for the average user, the reduction in latency and greater ability to multitask probably outweighs a few FPS in Doom 3.
My fan "repair" technique: remove backing (normally a sticker), spray with WD-40 (optional; just to clean it out if necessary), then put in a drop of oil, , make sure it's getting into the bearings, soak up excess, reattach sticker or seal with tape; done.
If a fan seems to be struggling it's a good idea to re-lubricate it like this before the bearings actually get damaged. I always find myself doing this on crappy chipset fans; it can easily double their lifespan.
VFS = also the kernel-layer filesystem API, on which everything from procfs to ext2 is interfaced. Bypassing this layer for certain operations on certain filesystems certainly sounds like something that should be tricky to get merged into mainline.
"I can't stand the rendering implementation in Opera, it's borked (yes seriously it has a lot more bugs than firefox"
:)
:)
This isn't my experience. Obscure clipping bugs, rendering problems and general clunkiness seem to be mainly the realm of IE and Gecko when I'm knocking out CSS and HTML. YMMV, a few anecdotes doesn't exactly make good evidence
Ironically when I look up half these bugs in Gecko, the QA contact is a guy who works for Opera QA..
"I'm not sure Opera even supports XSLT yet does it?"
No. Am I expected to believe this is a significant omission next to all the other stuff it supports? I'm sure if more than 3 demo websites try using it Opera will adapt. Does Firefox support SVG out of the box yet? Aural stylesheets and voice I/O? BitTorrent? I dare say these are more interesting to more people than an awful XML transformation language most people like purely server-side.
"(something that even IE has supported since version 5!)"
Right. Wasn't it based on working draft specifications and thus pretty much entirely broken by release? I remember it barely being able to parse well formed XML, never mind any languages using it.
I'm in the UK; even the cheapo £30 players Amazon do are generally multiregion, and disabling the region protection on supposedly fixed players seems generally a case of typing in a few sequences on the remote. Do you not even get that over there? Do your TV's at least generally support PAL?
Of course these days you could have used /rescue/sh, etc. Should really have been there since day 1...
:)
If your OS lacks such a thing, making one using busybox or so comes highly recommended, remembering to make it statically linked and immutable afterwards
Sure, but you'll need to make sure you have a government approved electrical pulse generator implanted into your hippocampus and motor cortex to prevent you forming long term memories or making any copies while doing so.
The region-free DVD player we've got downstairs was sold to us by Asda.. seems legal enough to me; it even happily ignores UOP's.
Oh, did you just mean all the ones in *your* country? Aren't you allowed to remove region coding on the basis of interoperability, BTW?
Please restate your opinion in the form of a rational argument, rather than using emotionally laden but ultimately meaningless words.
Opera already has this in the 8.10 beta: see their snapshots site. Seems to work nicely; it's not exactly Azureus, but it works, mostly.
Where's the lack of control? I don't like Adaptec's attitude with regard to aaccli, but that's their choice; I have the choice not to buy any more of their shitty RAID controllers. Ditto with WiFi cards; I have control over whether I'm happy to have my hardware work even though the software may be closed because those that make it are forced to make it so by dicks, or I can get something good.
Delivering "ultimatums" which everyone knows are going to be ignored just alienates people, users, developers and companies alike. Here in the real world, I need aac(4) to work because they're in a couple of important machines, and I wouldn't appreciate FreeBSD removing it just because the only online management tool is only available from Dell/Red Hat as a Linux binary. What does removing it gain anyone, aside from making Adaptec less likely to fund/support future FreeBSD driver development and pissing off those who actually use the cards?
My appologies for being so pragmatic; however, my attitude is more along the lines of "if company x wants to be lame and only provide binaries, fine, I'll just buy from company y who do provide open source drivers". Part of control is being able to do non-optimal things.