Re:Coming soon in version 10!
on
Opera 9.0 Released
·
· Score: 2, Insightful
One could equally ask why web browsers tend to include FTP clients. In practice, many people tend to use their web browsers to download stuff, and BitTorrent is an increasingly-popular way to download stuff.
Note also that Opera manages to do everything but the kitchen sink in less than 4MB, while Firefox's executable (on Windows) is 6.5MB before you even consider the multitude of XPCOM components, XUL documents, XBL bindings and JavaScript source files that make it actually work. I'll stick with Opera.:)
Re:OK, since you're so well informed..
on
Opera 9.0 Released
·
· Score: 4, Interesting
It varies. Web Forms 2.0 is open and specified at the URL linked in the original summary, though it isn't actually finalized yet. It's also designed to be backwards compatible, so there's no reason why you can't go ahead and use most of it on sites now and suffer no ill-effects. As for canvas, I believe it now works in Firefox, Safari and Opera but obviously not IE. SVG can be added to most browsers via a plugin if they don't support it already.
Certainly we're not going to be making use of most of these things tomorrow, but it's getting to the point where IE is the only one left that doesn't support them. Obviously that's a biggy, but the IE team has shown recently that they are willing to play nice by implementing everyone else's adaptation of their XMLHttpRequest object, so it's not inconcievable that they'd implement some of these other new toys if they prove useful.
Re:OMG!? "Opera-specific extensions"!?
on
Opera 9.0 Released
·
· Score: 5, Insightful
I didn't want to bloat my summary by going into too much detail about that, but to be specific they've created a new rendering context for canvas which is designed to make it easier to create 2D games by giving more raw access to the framebuffer. It is using the designed-in extensibility for canvas (which was, of course, a Safari extension to HTML in the first place!) and Opera is working with other browser manufacturers on a 3D rendering context for canvas which will allow full hardware-accelerated 3D when it's done.
It's becoming increasingly vaugue these days what constitutes a "standard" in the web sphere. Various other organisations are springing up outside of the W3C and proposing their own extensions and new specs, and I for one am quite enjoying the new stuff we're seeing as a result of this "competition". Other such third-party "extensions" include XMLHttpRequest (Microsoft), canvas (Apple), opera-2dgame (Opera), Web Forms 2.0 (WHATWG) and probably other stuff I'm forgetting. Opera supports all this stuff and also supports several W3C standards to boot!
Re:A bit torrent client?
on
Opera 9.0 Released
·
· Score: 4, Informative
The main Opera.exe plus the Opera.dll that contains all the fun stuff still only adds up to a paltry 3.12MB (Windows version, obviously) even with all this stuff. It might be experiencing a bit of creeping featurism, but it doesn't seem to be suffering for it. I've noticed no speed decrease from Opera 8.51.
I'm actually quite pleased with the BitTorrent support; There have been many occasions when I've gone to download something and a site has offered both BitTorrent and a normal HTTP download, and I've picked HTTP just because it saves me launching some other app. Obviously the prolific downloaders aren't going to use it in preference to Azureus or uTorrent, but I expect it'd come in handy for more casual users and is also a good first step to greater adoption of BitTorrent.
A bunch of friends and I used to run a bunch of MegaHAL bots on an IRC network. A couple of them ran for several years. We let them talk to one another on channels sometimes, with appropriate rate-limiting. After a while the longer-running ones started to seem more and more insane as their databases grew larger and larger. Eventually one of them exploded and corrupted its database somehow; we couldn't be bothered to fix it, but it was a fun experiment while it lasted.
Linux could already "serve files" in the kernel
on
Linux 2.6.17 Released
·
· Score: 1
sendfile already allowed in-kernel copying from a disk file to an arbitrary file descriptor; Apache does this when serving static files. splice seems to generalize this to allow this transfer to be between two arbitrary file descriptors, which would be useful for things like load balancers that just want to make the initial connection to the backend node and forget about it, since they can now just splice two sockets and have done with it.
I use Postfix. It's simple to set up in a simple configuration where it relays outgoing mail to a smarthost and delivers incoming mail locally. If you are doing something more complex than that then I have no idea, since that's all I use it for, alongside Dovecot for IMAP access to the local mailboxes.
Having said that, I'm sure Exim is just as easy for a simple config like that. I really just use Postfix because I've always used Postfix.:)
The other flaw is that nearby users are able to recieve traffic that's not intended for them. This can be mitigated by whispering, but this tends to result in increased data corruption and the need for retransmissions.
I can confirm that integrating with Active Directory is easy. We use Wildfire at my office and it took me about five minutes to get it talking to the LDAP service on our domain controller. It then take a few more minutes to get it providing VCards based on the directory information. It worked like a dream.
We have users here using all sorts of different clients. I use Psi, while our single Mac user uses iChat. Some guys use Miranda IM so that they can use MSN Messenger as well. One lovely thing about Jabber is that it can support multiple connected clients for the same user too, meaning that when people are working from home (which happens quite a lot at my company) they can log into the Jabber server without disturbing their in-office client and chat with everyone as normal.
I'm sure it'll happen eventually. Telewest has been going broke for a while now. So many of their franchise areas still don't have digital services because they can't afford to upgrade, and people leave them because they don't offer digital services, thus creating a vicious cycle of decline. A merger with NTL (or more likely, NTL just buying Telewest outright) seems very likely at this point.
That said, I am operating on information from a couple of years back, since I used to live in one of the aforementioned franchise areas and have Telewest service. I now live in an NTL area, though, and I don't subscribe to any cable services.
It's that time of year where lots of people find themselves constantly sniffling and sneezing. I'm not a hayfever sufferer myself, but I know plenty of people who are. I was thinking just the other day that Hayfever is quite a ridiculous affliction, since it's essentially an intolerance of the outside world! Surely we should be adapted to live in the world we live in?
So while hayfever sufferers do have my sympathy at this time of year, I can't help but wonder where we went wrong that such a large number of people have bodies that are incapable of dealing with summer.
In addition, I think it's quite likely that the market for performances of music will remain, and (for now at least) the law grants artists a monopoly on performances of their own work, so it's likely to remain a viable business model for some time to come. This is, after all, often the only source of income for unsigned or less-known bands that aren't "big" enough to achieve shelf-space in the major music stores.
Back in the days of DOS applications, everyone had to roll their own UI because DOS wasn't going to do it for them. MS-DOS Editor used a completely different UI style to WordPerfect, which was completely different again to Neopaint. Every developer had to re-invent the UI wheel because the underlying API was so primitive.
We've come on in leaps and bounds since then. Operating systems provide complicated widgets like tree views as standard, and applications generally have lots of things in common with one another as a result of using the same underlying components. Web application development seems like a big step backwards, as everyone is once again rolling their own support for basic things like drag and drop, expandable tree views, colour pickers...
There are some handy libraries out there that can cut out some of the work, but it's no substitute for the functionality being provided by the platform itself: every site is inconsistent with every other site, and every site that uses these outside libraries carries around with it a glut of extra code that must be downloaded by every user. XUL was a good attempt at providing a more application-centered markup language, but it's suffered from poor separation from Mozilla's guts, making it hard to implement compatibly elsewhere. However, I'd love to see good support for XBL so that at least our home-rolled widgets can be packaged up nicely into bindings, separating the content from the functionality. It'd be nice if we could have some of the good bits of XUL's layout model in CSS too; CSS is far too centred around presentation of text-based documents but isn't so hot for application interfaces.
I'll abstain from calling "Web 2.0" innovation until we get to the point where everyone isn't busy reinventing two decades of platform and UI design.
All text has poor readability at 96dpi, but (to my eye, at least) sans-serif is more readable than serif at screen resolutions, since they generally simplify better when crammed into 100 fist-sized pixels.
Perhaps my eyes/brain have just got accustomed to sans-serif type because I've been reading from screens most of my life. I think the correct answer to this debate is "it depends on the reader".
Alternative stylesheets were a nice idea on the part of the CSS people at W3C, but in practice browsers seem to be pretty poor at remembering your choice of stylesheet between pages or between visits to the same page, meaning that you either need to switch it every time you click a link or install some crazy browser extension that tricks the browser into using the right one.
The main limiter right now is that web designers are forced to use bitmap images which are inherently sized in pixels. Therefore in order to make the design sensible the font sizes need to be displayed in pixels too. If we ever manage to get generally-available vector graphics by some means designers will then be able to start specifying everything -- including graphics -- in relative sizes so that the absolute size of something can go back to being a user preference. Only then will this ideal of designers leaving the body text at 100% be feasible.
Of course, there will still be the need for bitmaps for things like photos, but ideally we'd be sending high-quality photographs to clients and having the browsers scale them to suit some relative unit. The limiters here are that data transfer speeds are still abysmal for lots of users (mobile device users especially) and that most browsers are terrible at scaling bitmaps. The former could be fixed by some mechanism which allows browsers to only retrieve the resolution they need, much like browsers are currently able to retrieve a stylesheet that's appropriate for their form factor.
I think for the purposes of running a 32-bit Linux distribution you'd do fine with server, since then you can just pretend it's another machine and connect running apps to the 64-bit X server for seamless, network-transparent desktop goodness. For running Windows, Workstation is probably less frustrating.
Of course, a VM just for running 32-bit apps is probably overkill. A chroot environment with the 32-bit version of Ubuntu in it should do the trick, since AMD64 Ubuntu is perfectly capable of running 32-bit processes, it just doesn't have the 32-bit libraries there for them to load. Using a chroot environment does have some annoying consequences, but it has to be less annoying than simulating an entirely separate machine.
I tried using RunAs for a little while on my own machine, but I soon found the problem with tying access permissions to user accounts: when I run things as this other account -- "NurgledAdmin", say -- any settings/files the program creates are owned by that username rather than my main account.
What I really want is the ability to "Run As If I Were A Local Administrator" while still being my one, singular account. I guess Vista will allow this in the form of prompting before privileged operations much like MacOS does, but that doesn't help me much until Vista is actually released. "Run As" is not a magic bullet.
I disagree. You can get anything by paying for it if you have enough money. In this particular case, you can pay other people to, say, create you a spreadsheet. These are the two ways to approach every problem: Fix it yourself or pay someone else to fix it. I pay someone to clean my windows, and someone pays me to write software. One person can't do everything.
(and sure, I could clean my own windows, but I don't have a tall enough ladder and I'd rather be doing something else with my spare time.)
Improved hardware should enable us to do new and exciting things, not do the same old things in a more bloated and inefficient manner. I've yet to see anything on Vista's features list that falls into the "new and exciting" category, nor that should warrant such a high minimum spec.
I, for one, prefer my installers to be archives with a declarative install manifest rather than just-another-app. It would make me much happier if installers were limited in what they were able to do. Why on earth do all drivers come with installers these days? I remember a time when you installed a driver by inserting the install disk and telling Windows what directory the.inf file was in. They should have just dressed that UI up a bit rather than encouraging hardware vendors to bundle all manner of half-baked, broken "value add" applications along with their drivers.
My company generally releases betas that are complete in some respects but not others, in order to get feedback on those components that are complete. It's only when we hit "release candidate" stage that we are saying that the software is, as far as we are aware, ready to be the stable release should we get no further bug reports.
Never mind running Win98 software under XP. If you get hold of a copy of Windows 1.0 you can run the applications that came with that under Windows XP. The only quirk is that the app windows open at the smallest possible window size, because Windows 1.0 didn't support overlapping windows and so the apps didn't actually choose a size for themselves.
Microsoft's devotion to backwards-compatibility is astounding. It's just a shame that their architecture has to suffer because of it.
Slashdot Mirror
One could equally ask why web browsers tend to include FTP clients. In practice, many people tend to use their web browsers to download stuff, and BitTorrent is an increasingly-popular way to download stuff.
Note also that Opera manages to do everything but the kitchen sink in less than 4MB, while Firefox's executable (on Windows) is 6.5MB before you even consider the multitude of XPCOM components, XUL documents, XBL bindings and JavaScript source files that make it actually work. I'll stick with Opera. :)
It varies. Web Forms 2.0 is open and specified at the URL linked in the original summary, though it isn't actually finalized yet. It's also designed to be backwards compatible, so there's no reason why you can't go ahead and use most of it on sites now and suffer no ill-effects. As for canvas, I believe it now works in Firefox, Safari and Opera but obviously not IE. SVG can be added to most browsers via a plugin if they don't support it already.
Certainly we're not going to be making use of most of these things tomorrow, but it's getting to the point where IE is the only one left that doesn't support them. Obviously that's a biggy, but the IE team has shown recently that they are willing to play nice by implementing everyone else's adaptation of their XMLHttpRequest object, so it's not inconcievable that they'd implement some of these other new toys if they prove useful.
I didn't want to bloat my summary by going into too much detail about that, but to be specific they've created a new rendering context for canvas which is designed to make it easier to create 2D games by giving more raw access to the framebuffer. It is using the designed-in extensibility for canvas (which was, of course, a Safari extension to HTML in the first place!) and Opera is working with other browser manufacturers on a 3D rendering context for canvas which will allow full hardware-accelerated 3D when it's done.
It's becoming increasingly vaugue these days what constitutes a "standard" in the web sphere. Various other organisations are springing up outside of the W3C and proposing their own extensions and new specs, and I for one am quite enjoying the new stuff we're seeing as a result of this "competition". Other such third-party "extensions" include XMLHttpRequest (Microsoft), canvas (Apple), opera-2dgame (Opera), Web Forms 2.0 (WHATWG) and probably other stuff I'm forgetting. Opera supports all this stuff and also supports several W3C standards to boot!
The main Opera.exe plus the Opera.dll that contains all the fun stuff still only adds up to a paltry 3.12MB (Windows version, obviously) even with all this stuff. It might be experiencing a bit of creeping featurism, but it doesn't seem to be suffering for it. I've noticed no speed decrease from Opera 8.51.
I'm actually quite pleased with the BitTorrent support; There have been many occasions when I've gone to download something and a site has offered both BitTorrent and a normal HTTP download, and I've picked HTTP just because it saves me launching some other app. Obviously the prolific downloaders aren't going to use it in preference to Azureus or uTorrent, but I expect it'd come in handy for more casual users and is also a good first step to greater adoption of BitTorrent.
A bunch of friends and I used to run a bunch of MegaHAL bots on an IRC network. A couple of them ran for several years. We let them talk to one another on channels sometimes, with appropriate rate-limiting. After a while the longer-running ones started to seem more and more insane as their databases grew larger and larger. Eventually one of them exploded and corrupted its database somehow; we couldn't be bothered to fix it, but it was a fun experiment while it lasted.
sendfile already allowed in-kernel copying from a disk file to an arbitrary file descriptor; Apache does this when serving static files. splice seems to generalize this to allow this transfer to be between two arbitrary file descriptors, which would be useful for things like load balancers that just want to make the initial connection to the backend node and forget about it, since they can now just splice two sockets and have done with it.
I use Postfix. It's simple to set up in a simple configuration where it relays outgoing mail to a smarthost and delivers incoming mail locally. If you are doing something more complex than that then I have no idea, since that's all I use it for, alongside Dovecot for IMAP access to the local mailboxes.
Having said that, I'm sure Exim is just as easy for a simple config like that. I really just use Postfix because I've always used Postfix. :)
The other flaw is that nearby users are able to recieve traffic that's not intended for them. This can be mitigated by whispering, but this tends to result in increased data corruption and the need for retransmissions.
I can confirm that integrating with Active Directory is easy. We use Wildfire at my office and it took me about five minutes to get it talking to the LDAP service on our domain controller. It then take a few more minutes to get it providing VCards based on the directory information. It worked like a dream.
We have users here using all sorts of different clients. I use Psi, while our single Mac user uses iChat. Some guys use Miranda IM so that they can use MSN Messenger as well. One lovely thing about Jabber is that it can support multiple connected clients for the same user too, meaning that when people are working from home (which happens quite a lot at my company) they can log into the Jabber server without disturbing their in-office client and chat with everyone as normal.
I'm sure it'll happen eventually. Telewest has been going broke for a while now. So many of their franchise areas still don't have digital services because they can't afford to upgrade, and people leave them because they don't offer digital services, thus creating a vicious cycle of decline. A merger with NTL (or more likely, NTL just buying Telewest outright) seems very likely at this point.
That said, I am operating on information from a couple of years back, since I used to live in one of the aforementioned franchise areas and have Telewest service. I now live in an NTL area, though, and I don't subscribe to any cable services.
You do realise that NTL can already provide you with Telephone, TV and Internet service all down one line, right? :)
It's that time of year where lots of people find themselves constantly sniffling and sneezing. I'm not a hayfever sufferer myself, but I know plenty of people who are. I was thinking just the other day that Hayfever is quite a ridiculous affliction, since it's essentially an intolerance of the outside world! Surely we should be adapted to live in the world we live in?
So while hayfever sufferers do have my sympathy at this time of year, I can't help but wonder where we went wrong that such a large number of people have bodies that are incapable of dealing with summer.
In addition, I think it's quite likely that the market for performances of music will remain, and (for now at least) the law grants artists a monopoly on performances of their own work, so it's likely to remain a viable business model for some time to come. This is, after all, often the only source of income for unsigned or less-known bands that aren't "big" enough to achieve shelf-space in the major music stores.
Back in the days of DOS applications, everyone had to roll their own UI because DOS wasn't going to do it for them. MS-DOS Editor used a completely different UI style to WordPerfect, which was completely different again to Neopaint. Every developer had to re-invent the UI wheel because the underlying API was so primitive.
We've come on in leaps and bounds since then. Operating systems provide complicated widgets like tree views as standard, and applications generally have lots of things in common with one another as a result of using the same underlying components. Web application development seems like a big step backwards, as everyone is once again rolling their own support for basic things like drag and drop, expandable tree views, colour pickers...
There are some handy libraries out there that can cut out some of the work, but it's no substitute for the functionality being provided by the platform itself: every site is inconsistent with every other site, and every site that uses these outside libraries carries around with it a glut of extra code that must be downloaded by every user. XUL was a good attempt at providing a more application-centered markup language, but it's suffered from poor separation from Mozilla's guts, making it hard to implement compatibly elsewhere. However, I'd love to see good support for XBL so that at least our home-rolled widgets can be packaged up nicely into bindings, separating the content from the functionality. It'd be nice if we could have some of the good bits of XUL's layout model in CSS too; CSS is far too centred around presentation of text-based documents but isn't so hot for application interfaces.
I'll abstain from calling "Web 2.0" innovation until we get to the point where everyone isn't busy reinventing two decades of platform and UI design.
All text has poor readability at 96dpi, but (to my eye, at least) sans-serif is more readable than serif at screen resolutions, since they generally simplify better when crammed into 100 fist-sized pixels.
Perhaps my eyes/brain have just got accustomed to sans-serif type because I've been reading from screens most of my life. I think the correct answer to this debate is "it depends on the reader".
Alternative stylesheets were a nice idea on the part of the CSS people at W3C, but in practice browsers seem to be pretty poor at remembering your choice of stylesheet between pages or between visits to the same page, meaning that you either need to switch it every time you click a link or install some crazy browser extension that tricks the browser into using the right one.
The main limiter right now is that web designers are forced to use bitmap images which are inherently sized in pixels. Therefore in order to make the design sensible the font sizes need to be displayed in pixels too. If we ever manage to get generally-available vector graphics by some means designers will then be able to start specifying everything -- including graphics -- in relative sizes so that the absolute size of something can go back to being a user preference. Only then will this ideal of designers leaving the body text at 100% be feasible.
Of course, there will still be the need for bitmaps for things like photos, but ideally we'd be sending high-quality photographs to clients and having the browsers scale them to suit some relative unit. The limiters here are that data transfer speeds are still abysmal for lots of users (mobile device users especially) and that most browsers are terrible at scaling bitmaps. The former could be fixed by some mechanism which allows browsers to only retrieve the resolution they need, much like browsers are currently able to retrieve a stylesheet that's appropriate for their form factor.
Slashdot still has a "light" mode. I've used it for years and am using it right now. It's a user preference setting.
I think for the purposes of running a 32-bit Linux distribution you'd do fine with server, since then you can just pretend it's another machine and connect running apps to the 64-bit X server for seamless, network-transparent desktop goodness. For running Windows, Workstation is probably less frustrating.
Of course, a VM just for running 32-bit apps is probably overkill. A chroot environment with the 32-bit version of Ubuntu in it should do the trick, since AMD64 Ubuntu is perfectly capable of running 32-bit processes, it just doesn't have the 32-bit libraries there for them to load. Using a chroot environment does have some annoying consequences, but it has to be less annoying than simulating an entirely separate machine.
I tried using RunAs for a little while on my own machine, but I soon found the problem with tying access permissions to user accounts: when I run things as this other account -- "NurgledAdmin", say -- any settings/files the program creates are owned by that username rather than my main account.
What I really want is the ability to "Run As If I Were A Local Administrator" while still being my one, singular account. I guess Vista will allow this in the form of prompting before privileged operations much like MacOS does, but that doesn't help me much until Vista is actually released. "Run As" is not a magic bullet.
I disagree. You can get anything by paying for it if you have enough money. In this particular case, you can pay other people to, say, create you a spreadsheet. These are the two ways to approach every problem: Fix it yourself or pay someone else to fix it. I pay someone to clean my windows, and someone pays me to write software. One person can't do everything.
(and sure, I could clean my own windows, but I don't have a tall enough ladder and I'd rather be doing something else with my spare time.)
Improved hardware should enable us to do new and exciting things, not do the same old things in a more bloated and inefficient manner. I've yet to see anything on Vista's features list that falls into the "new and exciting" category, nor that should warrant such a high minimum spec.
I, for one, prefer my installers to be archives with a declarative install manifest rather than just-another-app. It would make me much happier if installers were limited in what they were able to do. Why on earth do all drivers come with installers these days? I remember a time when you installed a driver by inserting the install disk and telling Windows what directory the .inf file was in. They should have just dressed that UI up a bit rather than encouraging hardware vendors to bundle all manner of half-baked, broken "value add" applications along with their drivers.
My company generally releases betas that are complete in some respects but not others, in order to get feedback on those components that are complete. It's only when we hit "release candidate" stage that we are saying that the software is, as far as we are aware, ready to be the stable release should we get no further bug reports.
Never mind running Win98 software under XP. If you get hold of a copy of Windows 1.0 you can run the applications that came with that under Windows XP. The only quirk is that the app windows open at the smallest possible window size, because Windows 1.0 didn't support overlapping windows and so the apps didn't actually choose a size for themselves.
Microsoft's devotion to backwards-compatibility is astounding. It's just a shame that their architecture has to suffer because of it.