And I could go for a 107 degree office and warm water right now. I'd even agree to sit on the floor. Weather is ridiculous here this year. Or any other year for that matter.
You're doing it wrong. Here's a tutorial on how you drink beer: 1. Order a full mug 2. Drink a half of it in one go 3. Socialize, eat something 4. Drink a half of the remaining half 5. Socialize, eat something 6. Drink the remainder 7. Go to step 1
Yes they do. Simple test for you - install Ubuntu 8.10 from the stock CD, upgrade it with 300M of upgrades, and leave Firefox running while this is going on, try to use it afterwards.
Java is a flaming turd of a programming language compared to Scala. About time someone made it outdated. Hopefully other "fad" sites follow the trend and switch to Scala as their language of choice, and then serious dev houses follow. Frankly, after reading the Scala book, I can't see a good reason for a team with little to no legacy Java code to even bother with Java.
Programs often get borked when they're upgraded on the fly. Firefox certainly does this if you upgrade it while it's running, as do various pieces of server software which HAVE to undergo a restart to update the data and config files.
You have to restart everything that was affected to be sure (including the OS if you upgrade the kernel).
You're forgetting that ASP.NET is much younger. And Java COMPLETELY failed on the desktop and on the web. Not so with.NET, which is about to come to fruition with Silverlight and WPF.
I only hope that someone company with good management buys them out. There are very few of those, but they do exist.
Sun could have OWNED the entire server side, the way Microsoft owns the desktop, if only they played their Java deck of cards as well as Microsoft is playing.NET. Young uns don't remember it now, but there was a time when Microsoft was scared shitless of Java, and rightfully so. You install a runtime and the OS sorta doesn't matter anymore - that goes to the core of their entire strategy and rips it apart.
The problem was (and is) that Sun's software strategy was sorta like a chicken running with its head cut off - it went from the web to embedded to desktop to servers and everywhere in between without getting particularly good at anything (at least not thanks to Sun's efforts - community saved their server story, but that's about it).
What they should have done is they should have absolutely nailed desktop and server, and done so in late 90's before their cash cow hardware and support business started drying up.
McNealy is single handedly responsible for Sun's demise. Instead of building Java platform into a formidable weapon that would let them take over the world pretty much, he spent much of the late 90's trying to screw with Microsoft, when it wasn't even seriously in the enterprise server business - Sun's core market.
There was NOTHING Microsoft could do to stave off Java except for two things: 1. Brain dead reliance on bytecode interpreter in early Java VMs (compare that to unconditional JIT on first call in.NET). 2. McNealy's preoccupation with secondary issues, like keeping Java pure on MS platform. What he should have been thinking of is how to make it BETTER than MS implementation. Microsoft VM blew the doors off Sun's own at the time, its UI controls looked native (they WERE native), it had much faster startup time. The situation with lack of portability would have rectified itself had Sun's stack been superior to Microsoft's - people would just develop for Sun's version and ship a JRE on CDs, no big deal.
The only thing I want from them (or whoever buys them in the end) at this point is release ZFS under GPL. It's seriously difficult to get me excited with anything computer related these days, and ZFS is one of those things I want really bad on my Linux boxes (I know there's FUSE version, but I want production quality code).
After they do that, they can just fold up the tent and go out of business. I wouldn't care.
It does lack focus, though, and that's "by design". There are tons of projects that exist solely because someone has fun working on them, not because they're "strategic".
If I may, I'd like to suggest one project where I'd really like to see feature parity with Windows.
I want transparent access to virtual file systems from any console or GUI program. I.e. I should be able to do "cat cifs://computername/sharename/filename.txt" on the console, just like in Windows.
The problem with all these comparisons is that Apple's "low end" is equivalent to any PC manufacturer's "high end", and Apple's "high end" simply can't be compared to PCs in terms of quality, and turns out to be slightly cheaper than a "checklist equivalent" PC. What most checklists miss is that you're buying a well thought out machine built to last. PC manufacturers simply don't have enough attention to detail to realize that you might want two separate volume levels for your built in speakers and headphones, or that keyboard backlight requires not one but two light sensors to work properly, or that a lot of folks would prefer DVI or DisplayPort instead of D-SUB. They just repackage the same garbage as everyone else.
Once your wallet stops hurting from that 17" MBP, you will find it really hard to settle for anything less. It's "intangible" to others, but it's going to be very "tangible" to you.
SMB actually works pretty well. It's just that it, and other protocols (such as sftp, etc), needs to be WAY more integrated if Linux is to compete with Windows on the desktop. There needs to be some kind of "VFS interceptor" on top of everything else that would map the shares, log you in using single-sign-on (and therefore talk to Kerberos), etc, etc. It should be fairly low in the stack, so that console programs "just work" with it.
This is a very un-sexy bunch of work, so it won't get done. Right now file sharing on Linux (or Mac OS X for that matter) is a half assed, broken hack.
Transparent access to network shares from the command line and third-party apps. When I say "ls smb://servername/sharename" I want it to do something, just like it does in Windows. I want Java apps to be able to access unmounted shares, just like Gnome apps. I want Open Office to not ask me for a goddamn password again when I'm opening the file from a VFS SAMBA share. This is extremely amateurish and annoying.
Another thing that's partially broken is Active Directory interop. Granted, you can join AD these days fairly easily with Likewise, and things _sorta_ work. But on Ubuntu 8.10, for example (don't know about others) you won't be able to add a user to local groups anymore. WTF?
The point is, there are issues far more fundamental than differences in the UI. Community needs to focus on those, even though they're not as sexy.
>> if you want to work for Microsoft research, then that same resume is worthless
Although the majority of folks in MSR are researchers (which, in 95% of cases requires a PhD), not everyone there is a researcher. Developers are in no way restricted in terms of doing research themselves. Many of them publish papers do "tech transfers", etc.
And if you have a PhD in IT and a list of publications (if such a thing exists), they'll probably consider you for their "Datacenter Futures" group, particularly if you spend a couple of summers with them as a summer intern.
If he intends to stay in academia and get a PhD (or go to Google) and fancy-ass school on the resume will definitely help with employment, as will publications.
If it's something else, though, then yes, do what you like to do, just make sure you pick a reputable school.
In the end, none of this shit matters once you get your first decent job. What will matter is your network, references and reputation. No one will put you in a leadership position just because you took some classes in school five years ago.
If I were the OP, I'd forgo M.Sc. entirely unless he wants to stay in academia. And I'm an M.Sc. myself (CS/EE).
Try it in Vista and see for yourself. This is a MANDATORY feature in a business setting, where you might want to hook up a projector to your laptop every now and then.
So that GPL would have nothing to do with it. I'd just pay for the license, install my *.deb and be done with it. Heck, I'd even pay more than $50 for it, if it was a solid, comprehensive piece of software.
1) No, it's not "there". Ubuntu is the only popular distro that enables this stuff by default, and then only until they get sued by Apple and Microsoft. I want _legal_ support for this. 2) It's not a problem, if you're willing to license technologies. And paying $50 for the license is a heck of a lot cheaper than changing countries. 3). It's not "there". I have NVidia chip in my laptop and Ubuntu 8.10. NVidia drivers are installed. Out of the box, when I connect my external DVI display (through the docking station), nothing happens. If I go through NVidia config utility, I can get it to work, but then when I undock, my main display (on the laptop) doesn't become the only one, and half of my icons and windows are "shown" on the invisible second display. You can fix this by writing some scripts. My point is, I shouldn't have to. 4. Not quite. My Dell Latitude E6400 complains about soft reset of the hard drive every time I wake it up. If I were a "regular" user, I'd be scared to death by that error message, since I wouldn't know what it implies. Default power saving settings in Ubuntu make zero sense (and most "normal" people will never go there to change them).
I use Linux on my laptop, but even I have to agree.
What I want is a $50 add-on that will:
1. Fully and legally support bytecode interpreter and hinting for fonts. Bonus points for including decent fonts as well. 2. Support all major audio and video codecs. I shouldn't have to break any laws to get support for my digital media. Bonus points for not having to buy another codec pack when I upgrade my OS. 3. Support multi-monitor automatically when I connect a monitor (like Mac or Windows). 4. Work well on laptops. I should not see error messages about my hard drive failing to soft-reset every time I wake my laptop up from sleep.
I had to edit config files to get my laptop to recognize and bring up the second screen after docking. Editing config files should not be required for one of the core mobility scenarios.
Does this release fix the issue? I.e. if I just connect the monitor, will it get recognized automatically?
There are two distinct classes of this phenomenon:
1. People who just appear "brilliant" due to being good at obfuscation and self promotion. There are tons of these in any big company. They talk the good talk and move on shortly before management discovers they aren't all they're pretending they are.
2. People who ARE so brilliant, that bringing them to the lowest common denominator would be hugely unproductive and they'd probably just leave rather than comply. I've met a few folks like this when I was at Microsoft. It was a humbling experience.
So I say beat #1 into submission and let #2 do whatever the heck they want. The trick is to get your average front-line manager to figure out which is which. The distinctions that are obvious to a grunt in the trenches aren't all that obvious to them.
Slashdot Mirror
And I could go for a 107 degree office and warm water right now. I'd even agree to sit on the floor. Weather is ridiculous here this year. Or any other year for that matter.
You're doing it wrong. Here's a tutorial on how you drink beer:
1. Order a full mug
2. Drink a half of it in one go
3. Socialize, eat something
4. Drink a half of the remaining half
5. Socialize, eat something
6. Drink the remainder
7. Go to step 1
Yes they do. Simple test for you - install Ubuntu 8.10 from the stock CD, upgrade it with 300M of upgrades, and leave Firefox running while this is going on, try to use it afterwards.
>> It's completely transparent to me as user.
Not if you're using the daemon while it's being restarted, or rely on its state.
Java is a flaming turd of a programming language compared to Scala. About time someone made it outdated. Hopefully other "fad" sites follow the trend and switch to Scala as their language of choice, and then serious dev houses follow. Frankly, after reading the Scala book, I can't see a good reason for a team with little to no legacy Java code to even bother with Java.
Programs often get borked when they're upgraded on the fly. Firefox certainly does this if you upgrade it while it's running, as do various pieces of server software which HAVE to undergo a restart to update the data and config files.
You have to restart everything that was affected to be sure (including the OS if you upgrade the kernel).
OpenSolaris is a broken piece of shit with no community. Try to set it to boot in text mode, and you will see what I mean. :-)
>> ASP.NET and J2EE are about equal in the market
You're forgetting that ASP.NET is much younger. And Java COMPLETELY failed on the desktop and on the web. Not so with .NET, which is about to come to fruition with Silverlight and WPF.
I only hope that someone company with good management buys them out. There are very few of those, but they do exist.
Sun could have OWNED the entire server side, the way Microsoft owns the desktop, if only they played their Java deck of cards as well as Microsoft is playing .NET. Young uns don't remember it now, but there was a time when Microsoft was scared shitless of Java, and rightfully so. You install a runtime and the OS sorta doesn't matter anymore - that goes to the core of their entire strategy and rips it apart.
The problem was (and is) that Sun's software strategy was sorta like a chicken running with its head cut off - it went from the web to embedded to desktop to servers and everywhere in between without getting particularly good at anything (at least not thanks to Sun's efforts - community saved their server story, but that's about it).
What they should have done is they should have absolutely nailed desktop and server, and done so in late 90's before their cash cow hardware and support business started drying up.
McNealy is single handedly responsible for Sun's demise. Instead of building Java platform into a formidable weapon that would let them take over the world pretty much, he spent much of the late 90's trying to screw with Microsoft, when it wasn't even seriously in the enterprise server business - Sun's core market.
There was NOTHING Microsoft could do to stave off Java except for two things: .NET).
1. Brain dead reliance on bytecode interpreter in early Java VMs (compare that to unconditional JIT on first call in
2. McNealy's preoccupation with secondary issues, like keeping Java pure on MS platform. What he should have been thinking of is how to make it BETTER than MS implementation. Microsoft VM blew the doors off Sun's own at the time, its UI controls looked native (they WERE native), it had much faster startup time. The situation with lack of portability would have rectified itself had Sun's stack been superior to Microsoft's - people would just develop for Sun's version and ship a JRE on CDs, no big deal.
The only thing I want from them (or whoever buys them in the end) at this point is release ZFS under GPL. It's seriously difficult to get me excited with anything computer related these days, and ZFS is one of those things I want really bad on my Linux boxes (I know there's FUSE version, but I want production quality code).
After they do that, they can just fold up the tent and go out of business. I wouldn't care.
It does lack focus, though, and that's "by design". There are tons of projects that exist solely because someone has fun working on them, not because they're "strategic".
If I may, I'd like to suggest one project where I'd really like to see feature parity with Windows.
I want transparent access to virtual file systems from any console or GUI program. I.e. I should be able to do "cat cifs://computername/sharename/filename.txt" on the console, just like in Windows.
The problem with all these comparisons is that Apple's "low end" is equivalent to any PC manufacturer's "high end", and Apple's "high end" simply can't be compared to PCs in terms of quality, and turns out to be slightly cheaper than a "checklist equivalent" PC. What most checklists miss is that you're buying a well thought out machine built to last. PC manufacturers simply don't have enough attention to detail to realize that you might want two separate volume levels for your built in speakers and headphones, or that keyboard backlight requires not one but two light sensors to work properly, or that a lot of folks would prefer DVI or DisplayPort instead of D-SUB. They just repackage the same garbage as everyone else.
Once your wallet stops hurting from that 17" MBP, you will find it really hard to settle for anything less. It's "intangible" to others, but it's going to be very "tangible" to you.
Oh, and BTW, SMB IS a published protocol these days. Jeremy & Co are working on the next version of SAMBA straight from the specs.
SMB actually works pretty well. It's just that it, and other protocols (such as sftp, etc), needs to be WAY more integrated if Linux is to compete with Windows on the desktop. There needs to be some kind of "VFS interceptor" on top of everything else that would map the shares, log you in using single-sign-on (and therefore talk to Kerberos), etc, etc. It should be fairly low in the stack, so that console programs "just work" with it.
This is a very un-sexy bunch of work, so it won't get done. Right now file sharing on Linux (or Mac OS X for that matter) is a half assed, broken hack.
And glue it to the lid. There you go.
Transparent access to network shares from the command line and third-party apps. When I say "ls smb://servername/sharename" I want it to do something, just like it does in Windows. I want Java apps to be able to access unmounted shares, just like Gnome apps. I want Open Office to not ask me for a goddamn password again when I'm opening the file from a VFS SAMBA share. This is extremely amateurish and annoying.
Another thing that's partially broken is Active Directory interop. Granted, you can join AD these days fairly easily with Likewise, and things _sorta_ work. But on Ubuntu 8.10, for example (don't know about others) you won't be able to add a user to local groups anymore. WTF?
The point is, there are issues far more fundamental than differences in the UI. Community needs to focus on those, even though they're not as sexy.
>> if you want to work for Microsoft research, then that same resume is worthless
Although the majority of folks in MSR are researchers (which, in 95% of cases requires a PhD), not everyone there is a researcher. Developers are in no way restricted in terms of doing research themselves. Many of them publish papers do "tech transfers", etc.
And if you have a PhD in IT and a list of publications (if such a thing exists), they'll probably consider you for their "Datacenter Futures" group, particularly if you spend a couple of summers with them as a summer intern.
If he intends to stay in academia and get a PhD (or go to Google) and fancy-ass school on the resume will definitely help with employment, as will publications.
If it's something else, though, then yes, do what you like to do, just make sure you pick a reputable school.
In the end, none of this shit matters once you get your first decent job. What will matter is your network, references and reputation. No one will put you in a leadership position just because you took some classes in school five years ago.
If I were the OP, I'd forgo M.Sc. entirely unless he wants to stay in academia. And I'm an M.Sc. myself (CS/EE).
MS fonts are over 10 years old at this point. And they weren't that good to begin with.
Try it in Vista and see for yourself. This is a MANDATORY feature in a business setting, where you might want to hook up a projector to your laptop every now and then.
So that GPL would have nothing to do with it. I'd just pay for the license, install my *.deb and be done with it. Heck, I'd even pay more than $50 for it, if it was a solid, comprehensive piece of software.
1) No, it's not "there". Ubuntu is the only popular distro that enables this stuff by default, and then only until they get sued by Apple and Microsoft. I want _legal_ support for this.
2) It's not a problem, if you're willing to license technologies. And paying $50 for the license is a heck of a lot cheaper than changing countries.
3). It's not "there". I have NVidia chip in my laptop and Ubuntu 8.10. NVidia drivers are installed. Out of the box, when I connect my external DVI display (through the docking station), nothing happens. If I go through NVidia config utility, I can get it to work, but then when I undock, my main display (on the laptop) doesn't become the only one, and half of my icons and windows are "shown" on the invisible second display. You can fix this by writing some scripts. My point is, I shouldn't have to.
4. Not quite. My Dell Latitude E6400 complains about soft reset of the hard drive every time I wake it up. If I were a "regular" user, I'd be scared to death by that error message, since I wouldn't know what it implies. Default power saving settings in Ubuntu make zero sense (and most "normal" people will never go there to change them).
I use Linux on my laptop, but even I have to agree.
What I want is a $50 add-on that will:
1. Fully and legally support bytecode interpreter and hinting for fonts. Bonus points for including decent fonts as well.
2. Support all major audio and video codecs. I shouldn't have to break any laws to get support for my digital media. Bonus points for not having to buy another codec pack when I upgrade my OS.
3. Support multi-monitor automatically when I connect a monitor (like Mac or Windows).
4. Work well on laptops. I should not see error messages about my hard drive failing to soft-reset every time I wake my laptop up from sleep.
I had to edit config files to get my laptop to recognize and bring up the second screen after docking. Editing config files should not be required for one of the core mobility scenarios.
Does this release fix the issue? I.e. if I just connect the monitor, will it get recognized automatically?
There are two distinct classes of this phenomenon:
1. People who just appear "brilliant" due to being good at obfuscation and self promotion. There are tons of these in any big company. They talk the good talk and move on shortly before management discovers they aren't all they're pretending they are.
2. People who ARE so brilliant, that bringing them to the lowest common denominator would be hugely unproductive and they'd probably just leave rather than comply. I've met a few folks like this when I was at Microsoft. It was a humbling experience.
So I say beat #1 into submission and let #2 do whatever the heck they want. The trick is to get your average front-line manager to figure out which is which. The distinctions that are obvious to a grunt in the trenches aren't all that obvious to them.
Except MacPro has Nehalem Xeons which no one else has yet. A lot of people miss this little detail. :-)