Steve runs GNOME, Chris runs KDE. So now the user population is going to be split 50:50? Oh wait, you've never heard of them, but "Sweeney" is obviously so famous people will switch desktop environments just to be more like him.
Well, I share a little of this concern, but mostly it's OK because we have Linux-style stable vs. unstable series. If you want a WORKING version of Gimp, you have to live with last years feature set for now, just like 2.0.x versions of Linux didn't have nice 2.1.x features. Just like kernel 2.0.x, the stable Gimp version does get fixes during the life of the development version, but no additional functionality is planned in that series. Personally I hope the next stable Gimp can ship this year, but that's just my personal opinion.
Demo downloading now, probably won't play it myself, but I know someone who was desperate for this thing -- give it to them tomorrow (well, later today) Good mirrors, BTW -- 100K (kilo_bytes_!) per second to me in the UK
The only part of Gimp which isn't in the same league as PotatoShop is the colour-handling. For those of us who don't use dead trees in our work, that's no loss at all. I have the option here, if I like, to use PS4, but I prefer Gimp.
I think end-users will be presently surprised by the steps forward in 1.2 when we finally switch from crazy-feature-addition mode to bug fixing:)
PS Some bias expressed here, as a known Gimp developer:)
PPS If you don't like the interface, buy a mouse with the right number of buttons, and repeat after me "Easy to learn != Easy to use".
Trespass? Different rules where I come from. You can trespass on my land, and all I can do is tell you to leave, use minimum force to make you leave, and eventually call the police and have them remove you. In real space this rule makes a lot of sense (not least because property boundaries are less than obvious). If you don't want people wandering onto your land, build a damn fence. However, that said, the same rules can't be applied on the net. Trespassing in real-space uses up a lot of the intruders time, and there's no obvious advantage to it. I don't see any serious problems with individual trespass (mass trespass is illegal) where I live, but I do (did) have serious spam problems until our SMTP servers were configured to use RBL-like tactics.
The global Top Level Domains are big waste of all our time. The people crying out for ".firm" or ".amusing" or ".bookseller" are missing the point. Truely global classification systems don't work -- pick something arbitrary (and ISO country codes are definitely arbitrary) and then stick with it. The most succesful attempts at classifying general "stuff" are probably the library class mark schemes. However there are still at least two competing schemes, totally incompatible and they're both ARBITRARY. We could learn from that. If there had never been any gTLDs (not even *.INT, which is a farce anyway) we wouldn't be having this discussion. Making DNS a local issue in each geographical area would reduce the pressure (except maybe in the litigious US of A) on technical organisations to make political decisions.
You may be wondering why I show so much faith in national government. Do I really think they have our best interests at heart? Of course not. But I do know that they have poor co-ordination. If the UK government tries to stop the Fulchester Underwater Canoing Klubb (www.fuck.co.uk) they'll just spring up in.FI or.TW or who-knows-where. Making DNS law stop at geographic boundaries keeps both them and us happy.
Why does it seem stupid? The mobile companies had their entire allocation moved into 07xxx (some pre-1998 phones are still in 09xx until next year) because the 07xxx prefix was reserved for Personal Numbers. This means numbers which lead to a Person rather than a fixed geographic location. Right now it is just 2--3 button presses on my nk702 to redirect calls to my office BT line. No charge to me. No extra charge to the caller. If Orange can do this, FREE, today, why would it be hard for them to send my calls to One2One if I switch providers? OFTEL knows that mobile users are reluctant to change providers -- even for a better financial deal, if they have to tell 1 Zillion people their new number. So they made a reasonable request from Orange, Vodafone, etc. that they allow customers to use any 07xxx series number with their phone. Finally, this also avoids too much wasted allocation. Under the old system, if I bought a dual-band Nokia and then switched Orange --> Vodafone, and then Vodafone --> One2One, that would waste three numbers for several months, because the old ones can't be used until they've "gone cold". Now they just number the phone when I buy it and that number stays with me.
The ongoing death of the Amiga is an opportunity for most of us to see what we've got to look forward to when other architectures enter the later stages of their death. I would argue that the Amiga was the first architecture to (falsely of course) promise immortality. Unlike people who owned ZX Spectrums or Commodore 64s, which were inevitably dead ends, Amiga users believed that their next computer, and the one after that would both be Amigas too. In reality, ALL hardware architectures must die. Clever technical tricks can buy you an extra life (PowerPC Macs can run M68K code) but sooner or later your luck will run out. Bad management only accelerated this process. If you think the reaction of Amiga fans to the slow death of their favourite architecture is painful -- just wait for (1) The death of the Mac, as Apple struggle to bring themselves and their customers into the 21st century; and also (2) Death of the IBM-compatible PC. I predict that the PC will probably go first. The moaning and grumbling of Mac users when threatened with extinction will seem like NOTHING when compared with the wrath of PC users. Slashdot (or its future equivalent) will be besieged by people claiming that the 286 was a message from god, and that no-one really needs more than 4Gb of addressable memory anyway... Then, once ex-PC users have settled down to running PC/AT emulators on their ia64 machines, we've got the Mac Paradigm change. This could go two ways. Like PotatoShop users in general, they might say they really wanted the change all along, "Multiple undo was what I wanted anyway" == "Memory protection is a refreshing change"? Or they could go to the other extreme. Protestors tied to old PPC Mac displays; Articles in the paper calling for a boycott. All this is starting to come true already - the Intel-based SGI machines may run Linux and NT, but they certainly aren't "IBM-compatible". Apple are shipping an operating system with BSD inside it.
Nick Lamb "Linux isn't portable" -- Linus Torvalds
The endless "But I wanna get paid" whines are very annoying to people like me who actually (shock!) read the FSF position on this. If you want to be paid to write software, find someone who'll pay you to write Free software. Don't try to confuse Commercial with Proprietary. What are you being paid for anyway? To write MS Windows? Like most programmers you're probably getting to paid to write something because your company NEEDS that software. Even if it was Free instead of Proprietary, they'd still need it, and they'd still need to employ someone to write it. You. The GNU web pages make it perfectly clear that they neither expect nor desire a future where no-one is paid to write code. In fact when your skills are easily transferable, and customisation is the number one employer, there will likely be *more* well paid jobs, because more companies can use your talents. The spin on Free Software created by ESR's Open Source name doesn't affect the ability for you to get a job, it just changes the media perception (which is less relevant than ESR would like you to think). Business isn't looking at charities like the FSF anyway -- it's looking at Red Hat. So the business model for Free Software isn't the FSF's model, but the Red Hat model.
As someone already said (above) the file format specification IS public. This is easily as well documented as SMB-over-IP (implemented by Samba) and the various standards for Windows APIs (don't laugh, a lot of WINE was built on standards documents)
Even a court order won't force MS to write documents better than your average internal engineering specs, so we have as much as we're likely to get. The idea of asking that standards be written FIRST, is cool, but unlikely to make it into a settlement.
Gnumeric (the GNOME spreadsheet) will load many Excel files TODAY. If you download the latest stable tarball, and appropriate dependencies (you'll have almost everything if you run Gnome 1.0) you can try it yourself. http://www.gnome.org/gnumeric/
Done so far: Numbers, dates, text, formulae, references, simple style To be done: Better date-handling (sorry Mac users, no 1904 yet) and more special cases (what were those Wacky Excel guys thinking) + Lots more!
Gnumeric will probably *never* full reproduce the singing/dancing animated spreadsheet you got from Marketing, but it certainly will let you look at the price list which was wrongly sent to you in Excel format.
I'm told it saves Excel files too, but I don't have a use for that.
There are two kinds of "support" being meant here. Linux 2.2 will run on a 16 CPU machine, AFAIK, and it will happily report all the CPU#s in/proc/cpuinfo. If it doesn't actually *work* on a 16 CPU machine that's probably a bug you can expect to see fixed during 2.2.x
However, just because all 16 CPUs are working, doesn't mean the code was written with 16 CPUs in mind. Linux 2.2 scales poorly above say, 4 CPUs. The exact problems will vary depending on your application, but chances are that 16 CPUs won't buy you the performance it should.
NT has some of the same problems, but in W2K Microsoft say those problems are gone in the kernel architecture for up to 16 CPUs. For now, if you want to do SMP in a biiig way (not dual-processor) then you don't want INTEL INSIDE.
Nick.
Qt achieves parity with GTK+
on
qt 2.0 released
·
· Score: 1
The article mentioned above (about KDE 2.0 and KOffice) is still wrong. Both KDE 2.0 and GNOME 1.0 have a theme engine system. Both of them use theme support from a DSO, and include a DSO for pixmaps themes. The article would love you to think that KDE will be fast, because it doesn't use pixmaps. In reality, most theme authors are lazy, and most themes will use the pixmap engine. For GNOME there are already some clean, fast theme engines, and there's no reason to believe it will be any different for KDE 2.0 -- but that's still months away. Likewise there are slow, bloated themes for GNOME, as there will be for KDE.
It's also misleading (deliberately perhaps) for the KDE developers to talk about KDE and GNOME "standardising" on a drag-and-drop protocol. In reality, GTK+ (and therefore GNOME) supported as many DnD standards as possible, while KDE decided to go it alone. This is finally fixed by Qt 2.0. Try this, right now, in your Desktop Environment: Go into Netscape, and drag a URL icon onto the desktop or task bar. Doesn't work? Sorry, your desktop environment doesn't agree with "Be liberal in what you accept".
I won't address the KDE vs Gnome thing for the 100th time. Since this is about a Qt book though, I will say this:
C++ is a nasty hack. I learnt it, I used it, I got tired of it, and I'm still writing C, Scheme and Java. The limitations of C++ are demonstrated by the inclusion of "moc" in Qt. Since when did I need language extensions to program a GUI? The only other toolkit I'm familiar with that does this is... MFC, which is also based on C++
Maybe I jump to conclusions too easily, but I think C++ is a poor choice of language for an ostensibly platform-independent GUI. The language is still very immature, and doesn't even show OO techniques very well.
If you're determined that OO design is "it" then check out GTK+ _and_ Qt before you decide on one or the other based solely on the rhetoric of this book or the Trolls. If you're determined to use C++, check out GTK-- too.
Slashdot Mirror
Steve runs GNOME, Chris runs KDE.
So now the user population is going to be split 50:50?
Oh wait, you've never heard of them, but "Sweeney" is obviously so famous people will switch desktop environments just to be more like him.
Right. Troll boy.
Well, I share a little of this concern, but mostly it's OK because we have Linux-style stable vs. unstable series. If you want a WORKING version of Gimp, you have to live with last years feature set for now, just like 2.0.x versions of Linux didn't have nice 2.1.x features.
Just like kernel 2.0.x, the stable Gimp version does get fixes during the life of the development version, but no additional functionality is planned in that series. Personally I hope the next stable Gimp can ship this year, but that's just my personal opinion.
Nick.
Demo downloading now, probably won't play it myself, but I know someone who was desperate for this thing -- give it to them tomorrow (well, later today)
:)
:)
Good mirrors, BTW -- 100K (kilo_bytes_!) per second to me in the UK
The only part of Gimp which isn't in the same league as PotatoShop is the colour-handling. For those of us who don't use dead trees in our work, that's no loss at all. I have the option here, if I like, to use PS4, but I prefer Gimp.
I think end-users will be presently surprised by the steps forward in 1.2
when we finally switch from crazy-feature-addition mode to bug fixing
PS Some bias expressed here, as a known Gimp developer
PPS If you don't like the interface, buy a mouse with the right number of
buttons, and repeat after me "Easy to learn != Easy to use".
Nick.
Trespass?
Different rules where I come from. You can trespass on my land, and all I can do is tell you to leave, use minimum force to make you leave, and eventually call the police and have them remove you. In real space this rule makes a lot of sense (not least because property boundaries are less than obvious). If you don't want people wandering onto your land, build a damn fence.
However, that said, the same rules can't be applied on the net. Trespassing in real-space uses up a lot of the intruders time, and there's no obvious advantage to it. I don't see any serious problems with individual trespass (mass trespass is illegal) where I live, but I do (did) have serious spam problems until our SMTP servers were configured to use RBL-like tactics.
Nick.
The global Top Level Domains are big waste of all our time.
.FI or .TW or who-knows-where. Making DNS law stop at geographic boundaries keeps both them and us happy.
The people crying out for ".firm" or ".amusing" or ".bookseller" are missing the point. Truely global classification systems don't work -- pick something arbitrary (and ISO country codes are definitely arbitrary) and then stick with it.
The most succesful attempts at classifying general "stuff" are probably the library class mark schemes. However there are still at least two competing schemes, totally incompatible and they're both ARBITRARY. We could learn from that.
If there had never been any gTLDs (not even *.INT, which is a farce anyway) we wouldn't be having this discussion. Making DNS a local issue in each geographical area would reduce the pressure (except maybe in the litigious US of A) on technical organisations to make political decisions.
You may be wondering why I show so much faith in national government. Do I really think they have our best interests at heart? Of course not. But I do know that they have poor co-ordination. If the UK government tries to stop the Fulchester Underwater Canoing Klubb (www.fuck.co.uk) they'll just spring up in
Nick.
Why does it seem stupid? The mobile companies had their entire allocation moved into 07xxx (some pre-1998 phones are still in 09xx until next year) because the 07xxx prefix was reserved for Personal Numbers. This means numbers which lead to a Person rather than a fixed geographic location.
Right now it is just 2--3 button presses on my nk702 to redirect calls to my office BT line. No charge to me. No extra charge to the caller. If Orange can do this, FREE, today, why would it be hard for them to send my calls to One2One if I switch providers?
OFTEL knows that mobile users are reluctant to change providers -- even for a better financial deal, if they have to tell 1 Zillion people their new number. So they made a reasonable request from Orange, Vodafone, etc. that they allow customers to use any 07xxx series number with their phone.
Finally, this also avoids too much wasted allocation. Under the old system, if I bought a dual-band Nokia and then switched Orange --> Vodafone, and then Vodafone --> One2One, that would waste three numbers for several months, because the old ones can't be used until they've "gone cold". Now they just number the phone when I buy it and that number stays with me.
The ongoing death of the Amiga is an opportunity for most of us to see what we've got to look forward to when other architectures enter the later stages of their death.
I would argue that the Amiga was the first architecture to (falsely of course) promise immortality. Unlike people who owned ZX Spectrums or Commodore 64s, which were inevitably dead ends, Amiga users believed that their next computer, and the one after that would both be Amigas too.
In reality, ALL hardware architectures must die. Clever technical tricks can buy you an extra life (PowerPC Macs can run M68K code) but sooner or later your luck will run out. Bad management only accelerated this process.
If you think the reaction of Amiga fans to the slow death of their favourite architecture is painful -- just wait for (1) The death of the Mac, as Apple struggle to bring themselves and their customers into the 21st century; and also (2) Death of the IBM-compatible PC.
I predict that the PC will probably go first. The moaning and grumbling of Mac users when threatened with extinction will seem like NOTHING when compared with the wrath of PC users. Slashdot (or its future equivalent) will be besieged by people claiming that the 286 was a message from god, and that no-one really needs more than 4Gb of addressable memory anyway...
Then, once ex-PC users have settled down to running PC/AT emulators on their
ia64 machines, we've got the Mac Paradigm change. This could go two ways. Like PotatoShop users in general, they might say they really wanted the change all along, "Multiple undo was what I wanted anyway" == "Memory protection is a refreshing change"? Or they could go to the other extreme. Protestors tied to old PPC Mac displays; Articles in the paper calling for a boycott.
All this is starting to come true already - the Intel-based SGI machines may run Linux and NT, but they certainly aren't "IBM-compatible". Apple are shipping an operating system with BSD inside it.
Nick Lamb
"Linux isn't portable" -- Linus Torvalds
The endless "But I wanna get paid" whines are very annoying to people like me who actually (shock!) read the FSF position on this. If you want to be paid to write software, find someone who'll pay you to write Free software. Don't try to confuse Commercial with Proprietary.
What are you being paid for anyway? To write MS Windows? Like most programmers you're probably getting to paid to write something because your company NEEDS that software. Even if it was Free instead of Proprietary, they'd still need it, and they'd still need to employ someone to write it. You.
The GNU web pages make it perfectly clear that they neither expect nor desire a future where no-one is paid to write code. In fact when your skills are easily transferable, and customisation is the number one employer, there will likely be *more* well paid jobs, because more companies can use your talents.
The spin on Free Software created by ESR's Open Source name doesn't affect the ability for you to get a job, it just changes the media perception (which is less relevant than ESR would like you to think). Business isn't looking at charities like the FSF anyway -- it's looking at Red Hat. So the business model for Free Software isn't the FSF's model, but the Red Hat model.
Nick.
As someone already said (above) the file format specification IS public.
This is easily as well documented as SMB-over-IP (implemented by Samba) and the various standards for Windows APIs (don't laugh, a lot of WINE was built on standards documents)
Even a court order won't force MS to write documents better than your average internal engineering specs, so we have as much as we're likely to get.
The idea of asking that standards be written FIRST, is cool, but unlikely to make it into a settlement.
Gnumeric (the GNOME spreadsheet) will load many Excel files TODAY. If you
download the latest stable tarball, and appropriate dependencies (you'll
have almost everything if you run Gnome 1.0) you can try it yourself.
http://www.gnome.org/gnumeric/
Done so far: Numbers, dates, text, formulae, references, simple style
To be done: Better date-handling (sorry Mac users, no 1904 yet) and more
special cases (what were those Wacky Excel guys thinking) + Lots more!
Gnumeric will probably *never* full reproduce the singing/dancing animated
spreadsheet you got from Marketing, but it certainly will let you look at
the price list which was wrongly sent to you in Excel format.
I'm told it saves Excel files too, but I don't have a use for that.
Nick.
There are two kinds of "support" being meant here. /proc/cpuinfo. If it doesn't actually *work* on a 16 CPU machine that's probably a bug you can expect to see fixed during 2.2.x
Linux 2.2 will run on a 16 CPU machine, AFAIK, and it will happily report all the CPU#s in
However, just because all 16 CPUs are working, doesn't mean the code was written with 16 CPUs in mind. Linux 2.2 scales poorly above say, 4 CPUs. The exact problems will vary depending on your application, but chances are that 16 CPUs won't buy you the performance it should.
NT has some of the same problems, but in W2K Microsoft say those problems are gone in the kernel architecture for up to 16 CPUs. For now, if you want to do SMP in a biiig way (not dual-processor) then you don't want INTEL INSIDE.
Nick.
The article mentioned above (about KDE 2.0 and KOffice) is still wrong.
Both KDE 2.0 and GNOME 1.0 have a theme engine system. Both of them use theme support from a DSO, and include a DSO for pixmaps themes. The article would love you to think that KDE will be fast, because it doesn't use pixmaps. In reality, most theme authors are lazy, and most themes will use the pixmap engine.
For GNOME there are already some clean, fast theme engines, and there's no reason to believe it will be any different for KDE 2.0 -- but that's still months away. Likewise there are slow, bloated themes for GNOME, as there will be for KDE.
It's also misleading (deliberately perhaps) for the KDE developers to talk about KDE and GNOME "standardising" on a drag-and-drop protocol. In reality, GTK+ (and therefore GNOME) supported as many DnD standards as possible, while KDE decided to go it alone. This is finally fixed by Qt 2.0.
Try this, right now, in your Desktop Environment: Go into Netscape, and drag a URL icon onto the desktop or task bar. Doesn't work? Sorry, your desktop environment doesn't agree with "Be liberal in what you accept".
Nick.
I won't address the KDE vs Gnome thing for the 100th time. Since this is about a Qt book though, I will say this:
C++ is a nasty hack. I learnt it, I used it, I got tired of it, and I'm still writing C, Scheme and Java. The limitations of C++ are demonstrated by the inclusion of "moc" in Qt. Since when did I need language extensions to program a GUI? The only other toolkit I'm familiar with that does this is... MFC, which is also based on C++
Maybe I jump to conclusions too easily, but I think C++ is a poor choice of language for an ostensibly platform-independent GUI. The language is still very immature, and doesn't even show OO techniques very well.
If you're determined that OO design is "it" then check out GTK+ _and_ Qt before you decide on one or the other based solely on the rhetoric of this book or the Trolls. If you're determined to use C++, check out GTK-- too.
Disclaimer: I've used GTK+ since 1997