unfortunately that's a problem too - Microsoft still leads in the server space too (though not by nearly as nuch as the desktop). Now I see Linux making big inroads to this market, but MS makes a ton of money off their server offerings - not just all the Windows Server flavours, but also all the server-based frameworks and 'servers'. Ever seen the cost of crappy old Biztalk server? even crappier Sharepoint? (especially as you also need Office Pro to get Infopath that makes it halfway worthwhile). Application Server? Exchange! It'd make your wallet cry to just look at the numbers that businesses regularly pay!.
Also, the server space is where MS get to say they are 'enterprise' and 'TCO' and 'competant'. They like marketing and they like money, so they wouldn't want to see this area be given up to Linux without a fight.
Here's a link to 2006 year-end figures, take a look at the bottom and see how the client profits have gone compared to server. (client take s a $700m hit, server rises $500m over the 6 months. The full year figures were skewed by Vista, but they show a trend towards server growth and client decline)
I live in the UK, British Telecom laid the phone lines to everywhere ages ago and then started broadband ADSL down them. At that point we had deregulated telecom market where anyone could setup shop and offer phone services using BT's wires, and they happily moved to offering ADSL too.
As a result we have massive competition, all of them competing on price (as they all sell the same thing - BT's ADSL, often they're just BT resellers). Now we have the exchanges opened up to them too, so some are putting their own networking kit there in place of BTs DSLAMs and offering fastre speeds (usually 24mbps), but these are only going in at city locations. You can see already that the exchanges are upgraded only where these companies think it makes financial sense to do so.
BT isn't doing much to bring fibre to the home, they've migrated fibre between exchanges, but if they spent the £10bn+ to put fibre in all the streets too, the reseller companies could resell it just as well as B and they'd get al the profit (or basically, BT couldn't charge as much as they'd need to). As a result, we're a bit stuck - no-one can put fast networks in because someone else will make money off it.
That's roughly the way it is in the UK, I'm sure that BT would lay fible, if they could sell it separately, but the government says otherwise. Alternative ISPs are putting faster kit in, but this is generally a half-way house - they're not digging up the street to lay fibre, just replacing DSLAMs with faster ones, and then only in some cities.
yeah, but think about what happens if you win your class action lawsuit.
Suddenly bandwidth starts to cost the amount the ISP think you might use, rather than the amount they think you will use. You get broadband for $20 a month for 10Gig transfer? Imagine it jumping to $50 a month as the ISP charges you for the increased capacity they have to buy at peak time.
So the class action would be a disaster, but so would the pitchfork approach - the ISPs are in such a market that competition has reduced their prices (and increased their marketing) to the point where people are complaining about what they get. I don't think there is the capacity like everyone wants.
If you want to get your pitchforks out, the only place to take them is central government. The major telco won't roll out fibre to the home because they'll be paying a lot for this, that ISPs would then get to use for next to nothing (monopoly aren't allowed after all) so they wouldn't get a return for the major cost of it all - especially as hardly anyone would take it up (adsl is perfectly good enough for everyone, not enough would want faster - not the majority who surf the web a bit and download email) The only one who can do it is either the government paying for it as a social thing (like roadbuilding), or allowing the telcos to charge for putting it in. I think it should be a government paid-for system. It'd cost £10b to fibre-up the UK. We give a damn sight more than that to poxy banks whose directors can't even add up properly.
this only applies to OSS software (generally). Commercial stuff is always download this in tar or rpm format and follow the instructions.
If Linux became seriously popular with the home users, how many really would install DancingBunnies.rpm?
If not do that (because they've heard that running rpms is bad, and they should only get it from repositories) how many would 'drop dancing bunnies.repo into their/etc/yum.d directory and then type 'yum update' because the install instructions told them to? (I shudder to think about that).
If someone says "Windows is insecure", I hear "Yeah, damn right. Stupid n00bs and its all Bill Gates fault, stupid people".
If someone says "Linux is insec.." I hear "lalalalalala. I can't hear you. lalalalalala".
The problem is about usage patterns of the OS. Put the same person in front of any OS and they will get infected the same way they always did. As someone mentioned, bots generally send spam or steal financial info - well, there's nothing stopping this from happening in any app. Either you restrict users from doing things they consider normal (like downloading gadgets and toys, and opening their own files) or you have to accept that they will get infected, no matter which OS they use.
Sure, there are technical, tricky issues with.bash_profile (and a thousand other ones), and you can configure/fix them out of existence. But to get all of them pretty much means stopping someone from using their computer.
The answer is to educate users about security, which would be an ongoing task forever (as new exploits are discovered, new attack vectors invented). Or to try and fix the damage an infected machine can do. Eg. why aren't the defaults for emailing set to only allow 1 per minute, or why doesn't the software pop a dialog every time an email is sent? If either of these were implemented at a point closer to the network (rather than the user application) then we'd get significantly less spam from infected PCs.
Of course, its tricky to do. A firewall could do it, but they tend to be focussed on on-demand access - ie, it'll pop a message everytime an app wants to use the network, and you end up with people turning the messages off.
Hiding the file extension - meaningless from a security viewpoint. Users still download SmileyCentral icon packs and explicitly install them.
a lawyer threatening a lawsuit should you not pay up? Sounds just as fishy as this..........
FROM THE DESK OF MR,CHARLES BOSAH MUTUAL TRUST LAWYERS & COMPANY HEAD QUATERS BRANCH VICTORIA ISLAND, LAGOS-NIGERIA. DEAR FRIEND,
My name is DR CHARLES OBOSAH, INTERNAL AUDITOR MUTUAL TRUST LAWYERS & COMPANY. I am writing in respect of a fee which you are owing to a customer of our bank Mr.JONATHAN GHUNIAM.
Since the fee has been outstanding fdor some time now, we must regretfully commence action upon you according to the laws of our country which would see you personally liable for a sum of up to US$9..5m (Nine Million Five Hundred Thousand United States Dollars) or a jail sentence of up to 424 years (Four Hundred Twenty Four years)
On this note, I decided that the late payment of said fee is a small matter to you, and should you wish to enter into communications with me for payment of said sum, we can arrange to close this matter if you will deposit $400 into the account via Western Union to the account described within the next 7 days.
I will not fail to bring to your notice that this business is hitch free safe and legal. we have to hire an attorney who will protect you legally,and for benefit of doubt position you as the next of kin and beneficiary.that you should not entertain any fear as all modalities for fund transfer can be finalized as soon as possible.
When you receive this letter, kindly send me an e-mail on this mail box including your most confidential telephone/fax numbers and your address for quick communication.
For now it's up to the users to decide if and when XP "dies". yup. We have a customer who still uses NT4, so I guess that MS doesn't generally get what MS wants.
Absolutely. Things were easier in those days though, and he had a single hardware platform to work with, and he did based much of it on Tripos. I heard it took him 2 weeks.
But anyway, with Microsoft, they'll never develop anything worth squat while they have more program managers, project managers, and other bureaucrats demanding estimates, plans and completion updates all the time. If MS took a MS-research project and tidied up the edges, then that'd be the most likely way Windows 7 will appear.
well, MS is moving back to native code... notice how little.net stuff is in Office and the OS itself, and as you know MS loves to use the stuff it creates ("eating its own dogfood") then I can only surmise that MS developers have tried it, didn't like it (or it didn't come up to snuff) and dropped it. Note how they have not written a TPC benchmark in.NET, where they usually love telling everyone how wonderful their new stuff is.
I was surprised to hear from the Visual studio program manager (on his blog) that they will be releasing the Office Ribbon bar as a component for MFC, and not.NET! ("we have no plans in the immediate future".... etc etc)
So for Windows 7, I imagine.NET will in fact be a legacy API - MS will come up with a new one for the new 'clean' API, and.NET will be part of one of the virtual subsystems.
We'll see - I think its all vapourware at the moment, though I can't think MS would not use virtualisation technologies in the next release of Windows.
Crefit where credit is due indeed - although 'cyberspace' was coined as a term in one of gibson's books (I thought it was Neuromancer, but I may well be wrong there), they were all strongly influenced by John Shirley's short story "Wolves of the Plateau" (you can buy it in the anthology Heatseeker).
Yes, you will get five million schmucks producing five billion memory-leaking resource-hogging maintenance nightmares, but you have to remember that those corporate applications are created to make something easier, to save work, and they usually do that, despite being the horrors they are Tell that to someone who's just lost his Word document and see how much he agrees with you.
We do need low barriers for entry, but that is not created by making things easy. Its made by encouraging junior programmers to start off junior and become better. It doesn't have to be hard at first, not if you have work that is within your skills and any form of mentor (usually the more senior programmers who were junior themselves). When I started coding, I didn't think I would be allowed to write the company's entire runtime framework.. and I wasn't. There's no surprise there, and I got better. Others can become "elite" too, but only if we do not accept sub-standard work as good. Imagine if we had the same low-barrier attitude towards nuclear engineering:-)
Besides, those maintenance nightmares make more work, so much so that you can end up spending all your time fixing them and not creating anything new or useful at all.
No, we need to encourage programmers to become better themselves, to learn better practices, to improve themselves. There's a place for juniors, and that's making the step towards becoming seniors.
Seeing as you're counting the STL in that list of features, I trust you do the same for Java or C# - how many do they have if you count the innumerable library features they support?
Teaching lisp, prolog, simula etc at university is an excellent idea. Make them think.
As for.Net being better - tell the Princeton university DARPA team that C# doesn't have memory leaks. Read Chris Brumme's blog for all the grubby hacks that MS made in the CLR to make things work.
the thing is, if you tell a less-capable programmer that he doesn't have to worry about memory, he'll think christmas has come early and he will stop worrying about memory... as a result his code will become poorer, resource consumption will increase dramatically, performance will tank and scalability will be non-existent. And his code will become even less maintainable.
Now make him program in a language where he is punished for cocking things up, he will *have* to become a better programmer or nothing he does will get released.
Now, that's an ideal, in the real world poor code will still get shipped to customers regardless of how it was developed; but the code written in an 'easy' langauge will still contain bugs.
eg. I saw a recent question on our internal tech mailing list - someone asked why his code was failing to connect to the DB, he was told "oh you have to put gc.collect() in any loop tht connects to the DB to fix that".
You don't need to tech kids to write in C++ for them to learn how bad an idea lack of GC is. Just give them some existing code and tell them to chase buffer overruns and memory leaks You don't need to tech (sic) kids to write in C# for them to learn how bad an idea of GC is. Just give them some existing code and tell them to chase memory leaks and finalisation issues.
I agree too, to quote TFA: "Conversely, there is now a generation who is firmly convinced that a program is only well-designed if just about everything is part of a class hierarchy and just about every decision is delayed to run-time."
I'm not sure if the problem is bad education or the lazy coders who expect everything to be easy (ie done for them) so they don't have to really think about what they're doing.
Even MS's clever chaps have this problem - "lets have C# and GC so no-one need think about memeory ever again" they cried. Then they realised that objects are more than just memory so you do have to worry about destruction of the non-memory resources held by an object (eg file handle, etc). Then they realised they were getting problems writing code that interacted with the OS, so they introduced reference counting objects that they could put in their "deterministic" finalisation objects that they could put in their Garbage-collected objects.
or.. needs bittorrent. Don't wikileaks host very large documents on their site? surely transferring that load to everyone else makes sense, not only because it reduces the load but also spreads the actual documents.
well, if this is the case why have they still got the domain name in use? Take it away (or better, let all connections through and then ignore them - 'dumbass' mail admins will soon figure out why their email servers are processing all that mail so slowly all of a sudden.
They could just remove the url and let the connections go nowhere - those admins would then see logs filled with 'failed to connect' errors.
Hurting people like this is not in the best interests of the net or community projects as a whole. It shows another reason why my boss can claim Linux (and all OSS stuff) is poor, this would never happen with a company he paid for service.
I say Jeeves, cancel my engagements for the morning, Aunt Agatha has decided that I must refactor my code so the Drones Club annual 'ship without testing' party will have to wait.
Adversity strikes when one least welcomes it Sir.
She claims my code 'smells'. I'll have her know my code smells as spiffly as a, as a, well, as a whatnot Jeeves.
Indeed sir.
Yes, a whatnot. I check my code against the very latest coding practices, and sometimes I even run it through unit tests!
Admirable qualities in a coder, if I may say, Sir.
Yes you may Jeeves. Now. to work! beastly testing.
Sir, perhaps one could use some automated tool or other method of achieving the requisite level of quality desired.
You know Jeeves, you've hit it right on the head there. I'll get Bernie Smetherington-Smythe to do it, he's such a ghastly bore but, well, when it comes to code review testing, there's no-one that can cut the mustard quite like him. Zip the source up Jeeves, we're to go pay Bernie a visit.
Certainly Sir, but what if Aunt Agatha finds out?
Pish Jeeves, pish! The auditors won't be around for months, no-one'll be any the wiser, and I can go to the ship-without-testing party after all. Life just falls into place sometimes doesn't it Jeeves? After all, What could go wrong?
One thing I do like about GNOME is that they have a built-in emacs key-binding option, which I can't figure out how to get in KDE yeah, well KDE has a built-in vi key-binding option, so there.
It kinda depends how much it'll cost. At the moment we run Linux servers as VM hosts, simply because buying Windows 2003 to run VMware on would cost us £500 a pop. Linux has proven to be stable, and works well.
As for other OS guests on Hyper-V, that depends too... sometimes MS finds 'unsupported options' that basically mean your favourite OS doesn't work, similar to how Vista wouldn;t run on VMware for some really obscure reason. (until VMware worked around the issue, that is).
, but slashdot keeps going on about how its a bloated piece of crap that uses 2GB of RAM when idle
that's 'cos the Aero glass interface is written in.NET, so it keeps as much data in memory until it gets garbage collected. Its a new paradigm in programming (for windows) - use up all your RAM for the running application, until you need it for something else, and then start swapping.
Swap is a strange thing, its never used by what you want it used for - either you use up all the ram you have (as free ram is wasted ram) but then, when you run out it starts to grind as existing app data is written to disk. Alternatively, you tell it to use as little as possible, but then you find it aggressively swaps unused apps and data out all the time.
Don't forget that things like prefetch simply move grind time from on-demand to on-startup, the time spent grinding is not reduced, just made more annoying. Ever wondered why your box grinds away when you're doing nothing?
The only answer is to have more ram than you could possibly need. If I had 1Tb RAM, I'd expect most of it would be given to file caching and app data. Unfortunately, I expect it'll be used as a huge managed heap waiting to be garbage collected.
Slashdot Mirror
unfortunately that's a problem too - Microsoft still leads in the server space too (though not by nearly as nuch as the desktop). Now I see Linux making big inroads to this market, but MS makes a ton of money off their server offerings - not just all the Windows Server flavours, but also all the server-based frameworks and 'servers'. Ever seen the cost of crappy old Biztalk server? even crappier Sharepoint? (especially as you also need Office Pro to get Infopath that makes it halfway worthwhile). Application Server? Exchange! It'd make your wallet cry to just look at the numbers that businesses regularly pay!.
Also, the server space is where MS get to say they are 'enterprise' and 'TCO' and 'competant'. They like marketing and they like money, so they wouldn't want to see this area be given up to Linux without a fight.
Here's a link to 2006 year-end figures, take a look at the bottom and see how the client profits have gone compared to server. (client take s a $700m hit, server rises $500m over the 6 months. The full year figures were skewed by Vista, but they show a trend towards server growth and client decline)
troll? eh?
I live in the UK, British Telecom laid the phone lines to everywhere ages ago and then started broadband ADSL down them. At that point we had deregulated telecom market where anyone could setup shop and offer phone services using BT's wires, and they happily moved to offering ADSL too.
As a result we have massive competition, all of them competing on price (as they all sell the same thing - BT's ADSL, often they're just BT resellers). Now we have the exchanges opened up to them too, so some are putting their own networking kit there in place of BTs DSLAMs and offering fastre speeds (usually 24mbps), but these are only going in at city locations. You can see already that the exchanges are upgraded only where these companies think it makes financial sense to do so.
BT isn't doing much to bring fibre to the home, they've migrated fibre between exchanges, but if they spent the £10bn+ to put fibre in all the streets too, the reseller companies could resell it just as well as B and they'd get al the profit (or basically, BT couldn't charge as much as they'd need to). As a result, we're a bit stuck - no-one can put fast networks in because someone else will make money off it.
That's roughly the way it is in the UK, I'm sure that BT would lay fible, if they could sell it separately, but the government says otherwise. Alternative ISPs are putting faster kit in, but this is generally a half-way house - they're not digging up the street to lay fibre, just replacing DSLAMs with faster ones, and then only in some cities.
Here's a link for a bit more information and more links
yeah, but think about what happens if you win your class action lawsuit.
Suddenly bandwidth starts to cost the amount the ISP think you might use, rather than the amount they think you will use. You get broadband for $20 a month for 10Gig transfer? Imagine it jumping to $50 a month as the ISP charges you for the increased capacity they have to buy at peak time.
So the class action would be a disaster, but so would the pitchfork approach - the ISPs are in such a market that competition has reduced their prices (and increased their marketing) to the point where people are complaining about what they get. I don't think there is the capacity like everyone wants.
If you want to get your pitchforks out, the only place to take them is central government. The major telco won't roll out fibre to the home because they'll be paying a lot for this, that ISPs would then get to use for next to nothing (monopoly aren't allowed after all) so they wouldn't get a return for the major cost of it all - especially as hardly anyone would take it up (adsl is perfectly good enough for everyone, not enough would want faster - not the majority who surf the web a bit and download email)
The only one who can do it is either the government paying for it as a social thing (like roadbuilding), or allowing the telcos to charge for putting it in. I think it should be a government paid-for system. It'd cost £10b to fibre-up the UK. We give a damn sight more than that to poxy banks whose directors can't even add up properly.
this only applies to OSS software (generally). Commercial stuff is always download this in tar or rpm format and follow the instructions.
/etc/yum.d directory and then type 'yum update' because the install instructions told them to? (I shudder to think about that).
If Linux became seriously popular with the home users, how many really would install DancingBunnies.rpm?
If not do that (because they've heard that running rpms is bad, and they should only get it from repositories) how many would 'drop dancing bunnies.repo into their
The thing is, I hear this all the time.
.bash_profile (and a thousand other ones), and you can configure/fix them out of existence. But to get all of them pretty much means stopping someone from using their computer.
If someone says "Windows is insecure", I hear "Yeah, damn right. Stupid n00bs and its all Bill Gates fault, stupid people".
If someone says "Linux is insec.." I hear "lalalalalala. I can't hear you. lalalalalala".
The problem is about usage patterns of the OS. Put the same person in front of any OS and they will get infected the same way they always did. As someone mentioned, bots generally send spam or steal financial info - well, there's nothing stopping this from happening in any app. Either you restrict users from doing things they consider normal (like downloading gadgets and toys, and opening their own files) or you have to accept that they will get infected, no matter which OS they use.
Sure, there are technical, tricky issues with
The answer is to educate users about security, which would be an ongoing task forever (as new exploits are discovered, new attack vectors invented). Or to try and fix the damage an infected machine can do. Eg. why aren't the defaults for emailing set to only allow 1 per minute, or why doesn't the software pop a dialog every time an email is sent? If either of these were implemented at a point closer to the network (rather than the user application) then we'd get significantly less spam from infected PCs.
Of course, its tricky to do. A firewall could do it, but they tend to be focussed on on-demand access - ie, it'll pop a message everytime an app wants to use the network, and you end up with people turning the messages off.
Hiding the file extension - meaningless from a security viewpoint. Users still download SmileyCentral icon packs and explicitly install them.
a lawyer threatening a lawsuit should you not pay up? Sounds just as fishy as this..........
FROM THE DESK OF
MR,CHARLES BOSAH
MUTUAL TRUST LAWYERS & COMPANY
HEAD QUATERS BRANCH
VICTORIA ISLAND,
LAGOS-NIGERIA.
DEAR FRIEND,
My name is DR CHARLES OBOSAH, INTERNAL AUDITOR MUTUAL TRUST LAWYERS & COMPANY. I am writing in respect of a fee which you are owing to a customer of our bank Mr.JONATHAN GHUNIAM.
Since the fee has been outstanding fdor some time now, we must regretfully commence action upon you according to the laws of our country which would see you personally liable for a sum of up to US$9..5m (Nine Million Five Hundred Thousand United States Dollars) or a jail sentence of up to 424 years (Four Hundred Twenty Four years)
On this note, I decided that the late payment of said fee is a small matter to you, and should you wish to enter into communications with me for payment of said sum, we can arrange to close this matter if you will deposit $400 into the account via Western Union to the account described within the next 7 days.
I will not fail to bring to your notice that this business is hitch free safe and legal. we have to hire an attorney who will protect you legally,and for benefit of doubt position you as the next of kin and beneficiary.that you should not entertain any fear as all modalities for fund transfer can be finalized as soon as possible.
When you receive this letter, kindly send me an e-mail on this mail box including your most confidential telephone/fax numbers and your address for quick communication.
Your Friend
DR CHARLES OBOSAH
we only get anywhere step by step, inch by inch.
Asteroid mining is still some way off, but its a little bit closer now.
oy! some of the best games are still DOS-only.
(good job we have DOSBox to run them)
Absolutely. Things were easier in those days though, and he had a single hardware platform to work with, and he did based much of it on Tripos. I heard it took him 2 weeks.
But anyway, with Microsoft, they'll never develop anything worth squat while they have more program managers, project managers, and other bureaucrats demanding estimates, plans and completion updates all the time. If MS took a MS-research project and tidied up the edges, then that'd be the most likely way Windows 7 will appear.
well, MS is moving back to native code... notice how little .net stuff is in Office and the OS itself, and as you know MS loves to use the stuff it creates ("eating its own dogfood") then I can only surmise that MS developers have tried it, didn't like it (or it didn't come up to snuff) and dropped it. Note how they have not written a TPC benchmark in .NET, where they usually love telling everyone how wonderful their new stuff is.
.NET! ("we have no plans in the immediate future".... etc etc)
.NET will in fact be a legacy API - MS will come up with a new one for the new 'clean' API, and .NET will be part of one of the virtual subsystems.
I was surprised to hear from the Visual studio program manager (on his blog) that they will be releasing the Office Ribbon bar as a component for MFC, and not
So for Windows 7, I imagine
We'll see - I think its all vapourware at the moment, though I can't think MS would not use virtualisation technologies in the next release of Windows.
Crefit where credit is due indeed - although 'cyberspace' was coined as a term in one of gibson's books (I thought it was Neuromancer, but I may well be wrong there), they were all strongly influenced by John Shirley's short story "Wolves of the Plateau" (you can buy it in the anthology Heatseeker).
We do need low barriers for entry, but that is not created by making things easy. Its made by encouraging junior programmers to start off junior and become better. It doesn't have to be hard at first, not if you have work that is within your skills and any form of mentor (usually the more senior programmers who were junior themselves). When I started coding, I didn't think I would be allowed to write the company's entire runtime framework.. and I wasn't. There's no surprise there, and I got better. Others can become "elite" too, but only if we do not accept sub-standard work as good. Imagine if we had the same low-barrier attitude towards nuclear engineering
Besides, those maintenance nightmares make more work, so much so that you can end up spending all your time fixing them and not creating anything new or useful at all.
No, we need to encourage programmers to become better themselves, to learn better practices, to improve themselves. There's a place for juniors, and that's making the step towards becoming seniors.
Oh, I thought the new American spelling of Java was indeed 'C#'. Perhaps its just a Seattle dialect.
Seeing as you're counting the STL in that list of features, I trust you do the same for Java or C# - how many do they have if you count the innumerable library features they support?
Teaching lisp, prolog, simula etc at university is an excellent idea. Make them think.
.Net being better - tell the Princeton university DARPA team that C# doesn't have memory leaks. Read Chris Brumme's blog for all the grubby hacks that MS made in the CLR to make things work.
As for
the thing is, if you tell a less-capable programmer that he doesn't have to worry about memory, he'll think christmas has come early and he will stop worrying about memory... as a result his code will become poorer, resource consumption will increase dramatically, performance will tank and scalability will be non-existent. And his code will become even less maintainable.
Now make him program in a language where he is punished for cocking things up, he will *have* to become a better programmer or nothing he does will get released.
Now, that's an ideal, in the real world poor code will still get shipped to customers regardless of how it was developed; but the code written in an 'easy' langauge will still contain bugs.
eg. I saw a recent question on our internal tech mailing list - someone asked why his code was failing to connect to the DB, he was told "oh you have to put gc.collect() in any loop tht connects to the DB to fix that".
There, fixed that for you.
I agree too, to quote TFA: "Conversely, there is now a generation who is firmly convinced that a program is only well-designed if just about everything is part of a class hierarchy and just about every decision is delayed to run-time."
I'm not sure if the problem is bad education or the lazy coders who expect everything to be easy (ie done for them) so they don't have to really think about what they're doing.
Even MS's clever chaps have this problem - "lets have C# and GC so no-one need think about memeory ever again" they cried. Then they realised that objects are more than just memory so you do have to worry about destruction of the non-memory resources held by an object (eg file handle, etc). Then they realised they were getting problems writing code that interacted with the OS, so they introduced reference counting objects that they could put in their "deterministic" finalisation objects that they could put in their Garbage-collected objects.
or .. needs bittorrent. Don't wikileaks host very large documents on their site? surely transferring that load to everyone else makes sense, not only because it reduces the load but also spreads the actual documents.
ook. ook ook, ook. uhuhuh ooooooook. oook oook, ahhha, ook ook.
(trans: if enough of us work at it, we're bound to make the perfect driver)
well, if this is the case why have they still got the domain name in use? Take it away (or better, let all connections through and then ignore them - 'dumbass' mail admins will soon figure out why their email servers are processing all that mail so slowly all of a sudden.
They could just remove the url and let the connections go nowhere - those admins would then see logs filled with 'failed to connect' errors.
Hurting people like this is not in the best interests of the net or community projects as a whole. It shows another reason why my boss can claim Linux (and all OSS stuff) is poor, this would never happen with a company he paid for service.
I say Jeeves, cancel my engagements for the morning, Aunt Agatha has decided that I must refactor my code so the Drones Club annual 'ship without testing' party will have to wait.
Adversity strikes when one least welcomes it Sir.
She claims my code 'smells'. I'll have her know my code smells as spiffly as a, as a, well, as a whatnot Jeeves.
Indeed sir.
Yes, a whatnot. I check my code against the very latest coding practices, and sometimes I even run it through unit tests!
Admirable qualities in a coder, if I may say, Sir.
Yes you may Jeeves. Now. to work! beastly testing.
Sir, perhaps one could use some automated tool or other method of achieving the requisite level of quality desired.
You know Jeeves, you've hit it right on the head there. I'll get Bernie Smetherington-Smythe to do it, he's such a ghastly bore but, well, when it comes to code review testing, there's no-one that can cut the mustard quite like him. Zip the source up Jeeves, we're to go pay Bernie a visit.
Certainly Sir, but what if Aunt Agatha finds out?
Pish Jeeves, pish! The auditors won't be around for months, no-one'll be any the wiser, and I can go to the ship-without-testing party after all. Life just falls into place sometimes doesn't it Jeeves? After all, What could go wrong?
Yes Sir.
umm..
That's probably not helping, is it.
It kinda depends how much it'll cost. At the moment we run Linux servers as VM hosts, simply because buying Windows 2003 to run VMware on would cost us £500 a pop. Linux has proven to be stable, and works well.
As for other OS guests on Hyper-V, that depends too... sometimes MS finds 'unsupported options' that basically mean your favourite OS doesn't work, similar to how Vista wouldn;t run on VMware for some really obscure reason. (until VMware worked around the issue, that is).
, but slashdot keeps going on about how its a bloated piece of crap that uses 2GB of RAM when idle
.NET, so it keeps as much data in memory until it gets garbage collected. Its a new paradigm in programming (for windows) - use up all your RAM for the running application, until you need it for something else, and then start swapping.
that's 'cos the Aero glass interface is written in
Swap is a strange thing, its never used by what you want it used for - either you use up all the ram you have (as free ram is wasted ram) but then, when you run out it starts to grind as existing app data is written to disk. Alternatively, you tell it to use as little as possible, but then you find it aggressively swaps unused apps and data out all the time.
Don't forget that things like prefetch simply move grind time from on-demand to on-startup, the time spent grinding is not reduced, just made more annoying. Ever wondered why your box grinds away when you're doing nothing?
The only answer is to have more ram than you could possibly need. If I had 1Tb RAM, I'd expect most of it would be given to file caching and app data. Unfortunately, I expect it'll be used as a huge managed heap waiting to be garbage collected.