I worked for Yahoo! for a while along with the author of the O'Reilly Regular Expressions book, Jeffrey Friedel. He had just about every O'Reilly book, recieved gratis, thanks to his being one of their authors. So if you want EVERY O'Reilly book, just write one! Chris Wareham
See the subject line. Just because it's a proprietary product doesn't mean it uses a proprietary format. Although I doubt at the cost whether many people have used it... Chris Wareham
I don't know whether anyone will read a post this far down, but I was offline in Amsterdam all day so I've only just logged in.
I have however, something to say on the matter of the 'first computer' debate that may be of interest. As one other post (at least) has pointed out, a number of people were working on similar ideas that arguably resulted in the first computer.
A good analogy is the jet engine. Both Wittel and the Heinkel firm (Britain and Germany respectively) were working on similar ideas at similar times. Ultimately the first jet plane to fly was German, but it is questionable who came up with the idea first.
Can't we just agree that the computer like the jet engine, was brought to frution during wartime, and ultimately benefitted most of us.
There is no love lost between myself and America (as a culture and ideal, not as individuals), but I can't help feeling that this nationalist ``we invented it first'' mentality is fruitless...
Do any other members of the Gnome team at RHAD have books on the way? A complete book on ORBit would be fantastic, as the OMG C binding document seems impossible to find at their site. Chris Wareham
The two books I would recommend to accompany this book are:
Advanced Programming in the Unix Environment by (the late and much lamented) W. Richard Stevens
Unix Systems Programming for SVR4 by David A. Curry
The latter is a much underated O'Reilly book, that despite concentrating on one flavor of Unix is very useful for the Linux programmers out there.
A really introductory book for Linux programming is `Beginning Linux Programming' by the usually iffy Wrox Press. It has a some naff stuff in it, but enough good bits to far outweigh the fluff like CGI and HTML tutorials that these sorts of book insist on having.
I spent a week of my holiday learning automake, etc. from the info pages. It was frustrating, but only because I started with the autoconf info pages, and then moved onto the automake stuff after a detour through the m4 pages...
Using the Gnome help browser was a major help - I'm not an emacs person, so the standard info reader was very unintuitive for me. For anyone contemplating learning the auto* thing, then I can recommend reading the automake page first. For smaller applications that don't rely on unusual libraries this makes the whole thing painless.
For adding dependency checks (like looking for the GTK+ libraries), just cut and paste from existing Gnome and GTK+ apps. Only if your using some really of the wall librarues will you have to learn the m4 macro language.
... I still feel that the GTK+ library an Gnome framework needs tyhe O'Reilly treatment. An exhaustive volume on each GTK+ widget and a similar tome on the Gnome GUI libraries.
As for things like CORBA and XML as used in Gnome, these are probably worthy of a volume each themselves. As Havoc Pennington's book points out, the while Gnome environment is essentially various libraries that will make application development much easier.
Note that I used the future tense. For those of us with little time to learn from the source of existing apps, good books are a must. Perhaps what we need is an entire Gnome series in the spirit of the O'Reilly X Programming books.
Anyway, I'm now off to buy the printed copy of this book to show my appreciation that Havoc's publishers allowed him such an amazing license.
Nearly all the X programming series books grace my bookcase at home (including the XView ones...). But when are there going to be companion volumes on GTK+ and the Gnome libraries? Get David Flanagan et. al. on the case now. Please!
I don't think anyone can say what a good age to start programming is. I have seen C and C++ code written by people in their early teens that is simply inspired. On the other hand I have seen code written by adults that made me cringe...
Personally, I learnt the rudiments of programming on a Commodore 64 (Basic), and then C. This was between the ages of about 12 - 16. Then I didn't touch a computer again until the age of 22. Five years later I now consider myself to be a pretty good C, C++ and Perl programmer.
So I don't think age is the issue, but motivation. To a certain extent, an individual can learn quite difficult concepts as long as they have a burning desire to. In my case I didn't want to be stuck in a bank doing data entry for the rest of my working life - the need to always be learning something new is what motivated me.
... C isn't. While I don't want to knock Python, I feel it is more useful for a beginner to learn a fairly low level language like C first.
The problem I have noticed with CS graduates who have learnt high level langauges to the exclusion of all else, is that they can't then grasp `real world' programming concepts. This may only be a problem afflicting British CS grads, but it is becoming increasingly hard to find good staff.
Programming isn't easy, although languages like Perl and Python can make certain tasks much easier than they were with C. The problem is making the distinction between tasks that should be tackled with tools like Python, and those that should be tackled with C or C++.
Perhaps it is time that CS degrees were split into systems programming courses and scripting based ones. There is a massive demand for programmers who will rarely need to touch compiled languages like C (with its memory allocation and lack of inbuilt dynamic types). So differentiating courses this way could allow a student to focus on one of these two distinct areas.
Or maybe this is just a refelction of my disgust at the low skillsets of CS grads being produced nowadays...
The slightly outdated ICCCM specification was an attemt to make toolkits, clients and window managers (which are just specialised clients), communicate with each other. As the Blackbox author points out, ICCCM needs some revision to bring it into the late nineties, but it forms a good basis for ensuring X clients don't behave oddly. I don't know whether olwm (the Open Look Window Manager) was ICCCM commpliant, but it always amazed me that I could run Motif apps under it without any problems... Chris Wareham
With 0.16 just around the corner this is great news. I've been running the last developer snapshot for a while now, and it's never crashed once. I don't know if it has any leaks in it, but I haven't seen any performance degradation yet... And speaking of performance, for such a feature loaded bit of software it runs faster on my aging Sparc 5 than any other window manager I've tried - that's some optimisation! Chris Wareham
(you wont get the subject line unless you've seen 'Only Fools and Horses', an occasionally funny British sitcom).
I've actually had my eye on a rather nice 21" premium Sun monitor (I've only got one of the cheaper 17" ones at home). I honestly don't know what is going to happen to all the hardware, but I doubt they'll offer it for sale to the staff.
Which reminds me. The Goth/Industrial club I go to (the legendary Slimelight in Islington, London) has converted one of its unused rooms into a warehouse for just about every kind of computer kit. Vax's, Suns and hordes of old PC's. Strange.
When the Yahoo! founders tried Linux it was in it's infancy. It just didn't cut it, and they needed something better than their existing setup (which was a bunch of older Sun stuff if I recall correctly).
The legend at Yahoo! (where I used to work until recently) is that David Filo got given a copy of FreeBSD and slapped it into the CD-Rom drive of a spare PC. It installed first time, and they've been using it ever since.
Yahoo! also uses some high end Sun stuff, but the majority of the servers run FreeBSD. Rather than rely on the potential for true scalability and SMP on single boxes, they use massed ranks of smallish machines.
I hated FreeBSD - but not because of it's stability. As a server OS on high quality hardware it is pretty much unbeatable in the Intel world. As a workstation OS it sucks a large one. We were stuck with 2.2.8, which has little to recommend it. In comparison, the 3.x series is much closer to the convenience of Linux as a workstation OS.
As for the ports and packages system, it is crap. It fills your hard drive with badly configured applications and their source trees that often wont compile. In comparison the RPM system of RedHat is a dream.
So, the FreeBSD kernel may be technically superior than Linux (and I'd be reasonably surprised if it wasn't), but Linux has the edge in user friendliness.
Which brings me to the Daemon News article. So the guys init scripts wouldn't port from one Linux distro to another... that is blatant FUD as the state of Unix init scripts has been a bone of contention for more years than FreeBSD and Linux has existed.
Like many *BSD zelaots, he makes the mistake of confusing the Linux kernel with the Linux distributions. The code 'chucked' into a Linux distro has nothing to do with the quality of the kernel that runs it.
But only because I have an aging Sun workstation at home...
For portablility, NetBSD is the champion of the *BSD's. If you need prepackaged security tools, then OpenBSD is your obvious choice. And if you're running an Intel machine then FreeBSD performs the best of the BSD's on that platform.
As for NetBSD... well I've now switched to RedHat 6.0 on my Sparc. It runs faster simply because of the development choices made by the SparcLinux team. The NetBSD guys had to make the decision whether to go for out and out performance on each platform that they support, or opt for sheer portablility convenience. To optimise for each platform would have meant much more work, and much more platform specific code. The SparcLinux team went for performance, and have a kernel that outperforms NetBSD - and seems as stable.
I was fascinated to see that the US DOJ, which must be a hive of petty bureaucrats, uses a free OS like OpenBSD. Here at... no I better not name them... I can't get authorisation for a Unix development server. It's bloody ludicrous.
I'm rewriting two systems that currently run on Solaris. I was expected to do this from an NT workstation running Hummingbird Exceed and connecting to a server in Amsterdam.
Firstly, the development server no lnoger exists. Secondly, our net connection to Holland is diabolical.
So I replaced NT with Linux on the sly, and rewrote the system in two weeks. Now I need to test it...
The recently appointed IT manager has made one decision since taking up his post:
Duhhh... no more Unix development. Everyfing must run on Windows NT - it's the future.
This has meant two things for me, a recently arrived contractor:
1) The Unix sysadmins *all* left in disgust at the switch to NT
2) We have storerooms full of unused Sun equipment, but no one will sanction the setting up of a new Unix server
Sorry about the pointless griping, but I had to get it off my chest.
Athena widgets are the only easy to use widget set that is guaranteed to be on any X installation. So it forms a nice lowest common denominator.
The standard X clients tend to use Xaw, even xterm. If you can't stand the look and feel of standard Xaw and find Xaw3d is less than exciting, then try neXtaw - the NeXT look and feel grafted onto Xaw.
Simply compile neXtaw and symlink the Xaw libs to it. It's available from http://windowmaker.mezaway.org/toolkits/nextaw.htm l
Sun's decision to unbundle a compiler from their operating system was lousy, but probably driven by the fact that many people use Sun workstations for non-programming tasks. The first place I worked at used a massive Sun setup to drive an optical jukebox system and their print shop. Funnily enough, they are still using the pre-Solaris SunOS - although that's because the jukebox software is no longer supported and wont run on Solaris.
As for the switch to System V, there are the compatability libraries. I've even found that Solaris is more likely to have the BSD function calls (albeit via wrapper libs) than FreeBSD. If you don't beleive me, then try compiling an application to use BSD style regexps. On Solaris it happily compiles. On FreeBSD it's deprecated...
I'm amazed at how consistent SVR4 is, especially when comapred to BSD Unices, so I can easily forgive Sun's switch.
I agree with your comment that StarOffice is overweight. This is not down to feature-bloat, but the fact that StarDivsion saw fit to create their own widgets that largely emulate the Windows 9x look and feel. Perhaps there are sound reasons why they have done this (portability?), but a switch to a common, lightweight toolkit would be good.
I'm shouldn't advocate one toolkit over another, but Qt and GTK+ seem the obvious choices. Qt especially if they want portability. And that's coming from someone who personally prefers coding in GTK+.
If I recall correctly, StarOffice used to use Motif up untill version four.
The O'Reilly Python book is dire, and why it's so bloody thick is beyond me. Python also performs far too badly for heavyweight usage, while Perl for all it's (many) sins is quite nippy.
In fact, I still haven't worked out why some of my Perl code is still faster than my carefully crafted C programs that do the same thing. Maybe it's the masses of resources that the Perl interpreter chews up that give it an edge...
(The Monty Python fixation put me off Python as well...)
On the other hand, he noted that Slashdot thrives in an environment where people seek more and more fragmentation in their lives. "Instead of knowing something about a lot of things, we know a lot about a little.
So concentrtating on a few subjects rather than many is fragmentation? I don't think so some how...
Slashdot Mirror
I worked for Yahoo! for a while along with the author of the O'Reilly Regular Expressions book, Jeffrey Friedel. He had just about every O'Reilly book, recieved gratis, thanks to his being one of their authors. So if you want EVERY O'Reilly book, just write one!
Chris Wareham
See the subject line. Just because it's a proprietary product doesn't mean it uses a proprietary format. Although I doubt at the cost whether many people have used it ...
Chris Wareham
I doubt that Knuth will ever consider ACP to be finished, but I hope he completes the first edition of each remaining volume.
Chris Wareham
I don't know whether anyone will read a post this far down, but I was offline in Amsterdam all day so I've only just logged in.
...
I have however, something to say on the matter of the 'first computer' debate that may be of interest. As one other post (at least) has pointed out, a number of people were working on similar ideas that arguably resulted in the first computer.
A good analogy is the jet engine. Both Wittel and the Heinkel firm (Britain and Germany respectively) were working on similar ideas at similar times. Ultimately the first jet plane to fly was German, but it is questionable who came up with the idea first.
Can't we just agree that the computer like the jet engine, was brought to frution during wartime, and ultimately benefitted most of us.
There is no love lost between myself and America (as a culture and ideal, not as individuals), but I can't help feeling that this nationalist ``we invented it first'' mentality is fruitless
Chris Wareham
Do any other members of the Gnome team at RHAD have books on the way? A complete book on ORBit would be fantastic, as the OMG C binding document seems impossible to find at their site.
Chris Wareham
I learnt GTK+ from the tutorial packaged with GTK+ itself. While it's not exhaustive, it is pretty damned good.
Chris Wareham
The two books I would recommend to accompany this book are:
Advanced Programming in the Unix Environment
by (the late and much lamented) W. Richard Stevens
Unix Systems Programming for SVR4
by David A. Curry
The latter is a much underated O'Reilly book, that despite concentrating on one flavor of Unix is very useful for the Linux programmers out there.
A really introductory book for Linux programming is `Beginning Linux Programming' by the usually iffy Wrox Press. It has a some naff stuff in it, but enough good bits to far outweigh the fluff like CGI and HTML tutorials that these sorts of book insist on having.
Chris Wareham
I spent a week of my holiday learning automake, etc. from the info pages. It was frustrating, but only because I started with the autoconf info pages, and then moved onto the automake stuff after a detour through the m4 pages ...
Using the Gnome help browser was a major help - I'm not an emacs person, so the standard info reader was very unintuitive for me. For anyone contemplating learning the auto* thing, then I can recommend reading the automake page first. For smaller applications that don't rely on unusual libraries this makes the whole thing painless.
For adding dependency checks (like looking for the GTK+ libraries), just cut and paste from existing Gnome and GTK+ apps. Only if your using some really of the wall librarues will you have to learn the m4 macro language.
Chris Wareham
... I still feel that the GTK+ library an Gnome framework needs tyhe O'Reilly treatment. An exhaustive volume on each GTK+ widget and a similar tome on the Gnome GUI libraries.
As for things like CORBA and XML as used in Gnome, these are probably worthy of a volume each themselves. As Havoc Pennington's book points out, the while Gnome environment is essentially various libraries that will make application development much easier.
Note that I used the future tense. For those of us with little time to learn from the source of existing apps, good books are a must. Perhaps what we need is an entire Gnome series in the spirit of the O'Reilly X Programming books.
Anyway, I'm now off to buy the printed copy of this book to show my appreciation that Havoc's publishers allowed him such an amazing license.
Chris Wareham
Nearly all the X programming series books grace my bookcase at home (including the XView ones ...). But when are there going to be companion volumes on GTK+ and the Gnome libraries? Get David Flanagan et. al. on the case now. Please!
Chris Wareham
I don't think anyone can say what a good age to start programming is. I have seen C and C++ code written by people in their early teens that is simply inspired. On the other hand I have seen code written by adults that made me cringe ...
Personally, I learnt the rudiments of programming on a Commodore 64 (Basic), and then C. This was between the ages of about 12 - 16. Then I didn't touch a computer again until the age of 22. Five years later I now consider myself to be a pretty good C, C++ and Perl programmer.
So I don't think age is the issue, but motivation. To a certain extent, an individual can learn quite difficult concepts as long as they have a burning desire to. In my case I didn't want to be stuck in a bank doing data entry for the rest of my working life - the need to always be learning something new is what motivated me.
Chris Wareham
... C isn't. While I don't want to knock Python, I feel it is more useful for a beginner to learn a fairly low level language like C first.
...
The problem I have noticed with CS graduates who have learnt high level langauges to the exclusion of all else, is that they can't then grasp `real world' programming concepts. This may only be a problem afflicting British CS grads, but it is becoming increasingly hard to find good staff.
Programming isn't easy, although languages like Perl and Python can make certain tasks much easier than they were with C. The problem is making the distinction between tasks that should be tackled with tools like Python, and those that should be tackled with C or C++.
Perhaps it is time that CS degrees were split into systems programming courses and scripting based ones. There is a massive demand for programmers who will rarely need to touch compiled languages like C (with its memory allocation and lack of inbuilt dynamic types). So differentiating courses this way could allow a student to focus on one of these two distinct areas.
Or maybe this is just a refelction of my disgust at the low skillsets of CS grads being produced nowadays
Chris Wareham
The slightly outdated ICCCM specification was an attemt to make toolkits, clients and window managers (which are just specialised clients), communicate with each other. As the Blackbox author points out, ICCCM needs some revision to bring it into the late nineties, but it forms a good basis for ensuring X clients don't behave oddly. I don't know whether olwm (the Open Look Window Manager) was ICCCM commpliant, but it always amazed me that I could run Motif apps under it without any problems ...
Chris Wareham
With 0.16 just around the corner this is great news. I've been running the last developer snapshot for a while now, and it's never crashed once. I don't know if it has any leaks in it, but I haven't seen any performance degradation yet ... And speaking of performance, for such a feature loaded bit of software it runs faster on my aging Sparc 5 than any other window manager I've tried - that's some optimisation!
Chris Wareham
(you wont get the subject line unless you've seen 'Only Fools and Horses', an occasionally funny British sitcom).
I've actually had my eye on a rather nice 21" premium Sun monitor (I've only got one of the cheaper 17" ones at home). I honestly don't know what is going to happen to all the hardware, but I doubt they'll offer it for sale to the staff.
Which reminds me. The Goth/Industrial club I go to (the legendary Slimelight in Islington, London) has converted one of its unused rooms into a warehouse for just about every kind of computer kit. Vax's, Suns and hordes of old PC's. Strange.
Chris
Chris Wareham
When the Yahoo! founders tried Linux it was in it's infancy. It just didn't cut it, and they needed something better than their existing setup (which was a bunch of older Sun stuff if I recall correctly).
... that is blatant FUD as the state of Unix init scripts has been a bone of contention for more years than FreeBSD and Linux has existed.
The legend at Yahoo! (where I used to work until recently) is that David Filo got given a copy of FreeBSD and slapped it into the CD-Rom drive of a spare PC. It installed first time, and they've been using it ever since.
Yahoo! also uses some high end Sun stuff, but the majority of the servers run FreeBSD. Rather than rely on the potential for true scalability and SMP on single boxes, they use massed ranks of smallish machines.
I hated FreeBSD - but not because of it's stability. As a server OS on high quality hardware it is pretty much unbeatable in the Intel world. As a workstation OS it sucks a large one. We were stuck with 2.2.8, which has little to recommend it. In comparison, the 3.x series is much closer to the convenience of Linux as a workstation OS.
As for the ports and packages system, it is crap. It fills your hard drive with badly configured applications and their source trees that often wont compile. In comparison the RPM system of RedHat is a dream.
So, the FreeBSD kernel may be technically superior than Linux (and I'd be reasonably surprised if it wasn't), but Linux has the edge in user friendliness.
Which brings me to the Daemon News article. So the guys init scripts wouldn't port from one Linux distro to another
Like many *BSD zelaots, he makes the mistake of confusing the Linux kernel with the Linux distributions. The code 'chucked' into a Linux distro has nothing to do with the quality of the kernel that runs it.
Chris - a NetBSD, Linux and ex-FreeBSD user.
Chris Wareham
But only because I have an aging Sun workstation at home ...
... well I've now switched to RedHat 6.0 on my Sparc. It runs faster simply because of the development choices made by the SparcLinux team. The NetBSD guys had to make the decision whether to go for out and out performance on each platform that they support, or opt for sheer portablility convenience. To optimise for each platform would have meant much more work, and much more platform specific code. The SparcLinux team went for performance, and have a kernel that outperforms NetBSD - and seems as stable.
For portablility, NetBSD is the champion of the *BSD's. If you need prepackaged security tools, then OpenBSD is your obvious choice. And if you're running an Intel machine then FreeBSD performs the best of the BSD's on that platform.
As for NetBSD
Chris
Chris Wareham
I was fascinated to see that the US DOJ, which must be a hive of petty bureaucrats, uses a free OS like OpenBSD. Here at ... no I better not name them ... I can't get authorisation for a Unix development server. It's bloody ludicrous.
...
... no more Unix development. Everyfing must run on Windows NT - it's the future.
I'm rewriting two systems that currently run on Solaris. I was expected to do this from an NT workstation running Hummingbird Exceed and connecting to a server in Amsterdam.
Firstly, the development server no lnoger exists. Secondly, our net connection to Holland is diabolical.
So I replaced NT with Linux on the sly, and rewrote the system in two weeks. Now I need to test it
The recently appointed IT manager has made one decision since taking up his post:
Duhhh
This has meant two things for me, a recently arrived contractor:
1) The Unix sysadmins *all* left in disgust at the switch to NT
2) We have storerooms full of unused Sun equipment, but no one will sanction the setting up of a new Unix server
Sorry about the pointless griping, but I had to get it off my chest.
Chris
Chris Wareham
Athena widgets are the only easy to use widget set that is guaranteed to be on any X installation. So it forms a nice lowest common denominator.
m l
The standard X clients tend to use Xaw, even xterm. If you can't stand the look and feel of standard Xaw and find Xaw3d is less than exciting, then try neXtaw - the NeXT look and feel grafted onto Xaw.
Simply compile neXtaw and symlink the Xaw libs to it. It's available from http://windowmaker.mezaway.org/toolkits/nextaw.ht
Chris
Chris Wareham
Microsoft was notified early Monday morning (August 30, 1999)
Sunday morning according to the Swedish paper that broke the story that day. But then maybe they don't work weekends, the lucky sods.
Please note that no action on your part is necessary to take advantage of the updated Hotmail
Wow, thanks.
Chris
Chris Wareham
Sun's decision to unbundle a compiler from their operating system was lousy, but probably driven by the fact that many people use Sun workstations for non-programming tasks. The first place I worked at used a massive Sun setup to drive an optical jukebox system and their print shop. Funnily enough, they are still using the pre-Solaris SunOS - although that's because the jukebox software is no longer supported and wont run on Solaris.
...
As for the switch to System V, there are the compatability libraries. I've even found that Solaris is more likely to have the BSD function calls (albeit via wrapper libs) than FreeBSD. If you don't beleive me, then try compiling an application to use BSD style regexps. On Solaris it happily compiles. On FreeBSD it's deprecated
I'm amazed at how consistent SVR4 is, especially when comapred to BSD Unices, so I can easily forgive Sun's switch.
Chris Wareham
I agree with your comment that StarOffice is overweight. This is not down to feature-bloat, but the fact that StarDivsion saw fit to create their own widgets that largely emulate the Windows 9x look and feel. Perhaps there are sound reasons why they have done this (portability?), but a switch to a common, lightweight toolkit would be good.
I'm shouldn't advocate one toolkit over another, but Qt and GTK+ seem the obvious choices. Qt especially if they want portability. And that's coming from someone who personally prefers coding in GTK+.
If I recall correctly, StarOffice used to use Motif up untill version four.
Chris Wareham
The O'Reilly Python book is dire, and why it's so bloody thick is beyond me. Python also performs far too badly for heavyweight usage, while Perl for all it's (many) sins is quite nippy.
...
...)
In fact, I still haven't worked out why some of my Perl code is still faster than my carefully crafted C programs that do the same thing. Maybe it's the masses of resources that the Perl interpreter chews up that give it an edge
(The Monty Python fixation put me off Python as well
Chris Wareham
Most of the code behind the unravelling of DNA is written in Perl. And they even make it freelly available ...
Chris Wareham
On the other hand, he noted that Slashdot thrives in an environment where people seek more and more fragmentation in their lives. "Instead of knowing something about a lot of things, we know a lot about a little.
...
So concentrtating on a few subjects rather than many is fragmentation? I don't think so some how
Chris Wareham