And if this car had 100x the sensor and computational capacity of the minimum needed to avert crashing in to walls, then it's ok?
I think we've all done it before, clicked the 'x' button too many times, pressed save on the document you didn't want to keep and don't save on the one you do. It might be user stupidity, but it's avoidable user stupidity.
UK Online is going to be bringing 16mbit/sec uncapped for £30/month sometime in the summer. That's very close to what that is.
Also, BT has 99.6% of the UK population covered with basic DSL, which is the world's highest.
BT as a private company which takes zero funding from the government has rolled broadband out very well.
They are launching a nationwide 8mbit service later this autumn, and trailing ADSL2.
This is going to mean nearly all of the population can have high speed DSL, whereas in france only the big towns can have it. I don't think you'll see villages with less than 300 people in them enabled in france, but you will in the UK.
As for the UK bandwidth thing, bascially you have Bulldog and UK Online who have put their own equipment inside of the exchange, which seriously reduces their costs. Instead of paying BT £45/month line rental for a 2mbit line, they just pay £8/month to rent the basic copper. After they have the DSLAMs and fibre in the exchanges, their ongoing costs are virtually nil.
Also, we have seen BT introduce 'CBC' - capacity based charging, which sees the ISP choosing how much bandwidth they want, and paying for it that way. I think they pay £11/month to BT for the line at the full speed it can go to (currently 2mbit) and then an insane price for their central bandwidth (in 155mbit chunks). This means it's very profitable to cap but not to allow heavy users on.
Actually, the concept of 'saving' is ridiculous, especially text documents. I saw a computer advertised with a 300GB HDD.
When there is 300GB of space, why on earth isn't the computer saving _every time_ the document is changed, and storing those revisions in a CVS-style way (obviously not with a CVS style interface, perhaps a 'revisions' sidebar with all the major changes listed).
Think about this user who goes and buys a new desktop PC with 300GB of disk space. Assuming he doesn't get infested with adware and they don't install games, they are going to use at most, what, 5GB of space for their documents, pictures and music? 50GB maybe if they also edit video.
There should be no 'save' command in applications any more. There should simply be a 'title document' command and 'save a copy' for when the user needs to copy it to a different computer, floppy disk, whatever.
AOL UK provides a fairly decent no-limits, straight up PPPoA ADSL broadband connection, which you can use with any DSL router or modem -- you don't need their software anymore.
They also provide access to their email via IMAP4.
I wouldn't choose them, but they are extremely well priced in the UK broadband market and have a very good network (in terms of peering, latency and speeds) -- at least on the DSL side of things.
Nearly every other major UK ISP caps users. British Telecom for example has a 512k connection with a 1GB cap for hte same price that AOL does a unlimited one.
Also on the subject of UK Broadband news, UK Online have dropped the price of their 8mbit service to £29.99/month.. which isn't bad at all when you consider it's free activiation and a free 802.11g wireless DSL router.
The reason it will be patented is that if they don't, some other company can just go ahead and patent the same thing, and the only way to rectify it is a long and hard court case which will cost lots and lots of $$$.
The fact is that the OS is not a big price for a major company. When I can go on dell.com and order 10 2.4GHz Celeron machines with a copy of WinXP Pro for $349, it's not a big deal at all.
Let's say these are for secretarial use. 99% of secreteries know how to use Windows, Word and Outlook.
Let's say I also spend $200-$300 (a day basically) on a techinican to set up a group policy and install Firefox on all these machines. These machines now can't run.exe,.pif etc etc and Firefox means veryl ittle crap is going to come in from the web.
Looking at the Windows startup cost it's $349x10 + $300 = $3790.
Now let's see the Linux cost. I'm going to get a maximum of $50 off those Dell machines for chosing Home instead of Pro, I can't 'not have' Windows on it. So that makes it $299/machine. Let's say the cost of installing Linux on each of these is $0.
Now let's look at my army of typists. None of them know how to use Linux/GNOME, OpenOffice or Evolution. So I train them. I hire a training guy to come in for a day to give them a crash course on how to use Linux, and he charges me $200. However, I've got to pay my typists anyway, $100/ea for the day. So that's $100x10 + $200 = $1200.
Linux startup cost: $2990 + $1200 = $4190. Windows wins.
Now, this is probably a bad example, but training costs, which are not going to change for the short to medium term, are very expensive.
For many small businesses this is the situation they have, and it's even worse if they have specialist apps they need to run on Windows.
So saying 'Linux has TCO ownage on Windows' is a bit unfair. It's very much true (IMO) for servers and workstations. But for average 'business desktops' I don't think it is.
I'd love to agree with you, but sadly graphics and physics engines are huge money-makers.
The Doom3 engine over its lifetime will sell for many times more than the Doom3 game itself sold for. I'd guess Valve wants to try and do this also with Source.
Not to mention both of these games I belive use the Havok physics engine which is done by a completely different company - I think all Havok does _is_ physics engines. Suggesting they GPL their only money maker is not going to work sadly.
I think Open Source's best place is going to be the building blocks for the apps that run on top of them. For example, a web application which you deploy via Firefox running on Linux. Or a database app written on top of pgSQL or mySQL. This is going to be a two way process - the people developing the 'front ends' to these blocks are going to be improving the back-end - it's in their interest.
It must of took, what, 3 years at least to even start making money with them. Now they have a market cap of over $40billion.
But yes, your advise is sound, but don't forget some of the best companies have been a complete gamble. Keeping costs low is always very important and if you can save $800 on your server setup with OSS, that could be the extra week than you need to survive in business and get that big contract.
I agree -- Python is fantastic for quickly building small apps, or even much larger ones.
The problem arises in Python's web programming support. The documentation is pretty much non-existent and you can soon get module-overload when you are importing more and more modules to do fairly simple stuff in web apps.
Sometimes I just think while Python is most certainly a far better designed language, PHP/ASP.NET (C#) seems much more 'pratical', and it's definitely much easier to quickly build web apps in.
Is there much effort to improve Python's web support? A manual with similar completion of php.net would help it go a lot lot further.
I think they mean more selling chip layouts to other companies for manufacture.
This is going to be a huge industry. If they could produce, say, a 800MHz CPU which ran on 1W or 0.5W of power and had sensible float performance, it could easily sell exceedingly well.
There must be millions upon millions of devices that require more than just PIC-level performance but low power consumption. Things like digital TV decoders -- the video itself can be decoded with a seperate chip but the amount of interactivity that will be delivered in the future is going to be immense.
Ok. I take what you say onboard. I do like what you say about automator and spotlight, and to be honest maybe I was not giving enough thought to them (especially automator, from Apple's screenshots it looks to me to be little more than some terribly limited graphical applescript).
Spotlight sounds intresting. I have used somewhat similar apps like Quicksilver (and some other one I can't think of) and it promised the same and failed horribly. Inaccurate, slow and unresponsive. So,
I really think that Apple is pushing the hardware _too_ hard. I think it will be amazing for someone running it on a dual core Powermac G5 3.0GHz. But for the millions (the majority) who are running on G3 iBooks or G4 iMacs/Powerbooks/eMacs/iBooks then I think it's going to be a pretty bad experience.
Sure, it sounds very 'science-fiction cool' to have the system 'automatically' choose the best system and reoptimise it for it, but it's really not very practical.
The fact is that the GPU will be 10-25x+ faster than Altivec vector processing unit. Altivec will probably be 2-5x faster than the CPU. If this is going to mean you get 5fps on an effect vs 100fps or more it's going to be a big deal. If you can only get 5fps with the vast majority of the userbase you are not going to spend loads of extra programming time and money creating a completetly different codepath for pixel shaders vs altivec vs normal, and scaling down the effects as you go along.
The thing is that Apple could put a NVidia 5200 on it and it would be able to do this. But instead they continue shipping the frankly bollocks Radeon 9200. How much would it cost to upgrade it? $10, $15?
Yes, but with the G4 the FSB is choked to hell already. Chucking down really complex pixel shading effects to do on the _CPU_ is going to slow it down so much it's going to be unusable for anything more than the most basic effects. If it was easy and fast to do on the CPU we would of already seen it. The fact is that you need a dedicated unit for this, on the GPU.
I agree entirely on your second point, apart from the time frame. I'd guess it'll be 3 years before the majority of the Mac userbase has a pixel shader onboard. I mean, after 5 years of OSX only 60-80% of the userbase has moved, and all that takes is a CD inserted into the drive, and not a new Mac that many will need for PS support.
I'm sure some of the new features will be useful, but to me it seems that Tiger is being completely overhyped.
Sorry, but for most people CoreImage and CoreVideo is going to be utterly useless. Apple still ships shit, shit, shit video processors on the iBook, Mac Mini and only the latest generation Powerbooks, PMs and iMac have the much-needed Pixel Shader on their GPUs. I'd guess probably 10-20% of the Mac userbase uses a Powerbook latest revision, PM G5, or iMac G5. The iBook was Apple's best selling Mac a few months back and I'm sure that the Mac Mini will replace it.
So are you honestly going to tell me developers are going to bother developing with features that only 10-20% of their already small userbase can use?
Personally I don't see any one feature that Tiger has that I really want. Hopefully it'll be a lot more polished and have some nice performance increases, but the vast majoirty of stuff in Tiger is totally useless to me: I don't need spotlight since I organise my stuff well, I don't use Safari for anything more than basic browsing (I have a perfectly good RSS client already, thanks), I won't be using automator, quicktime or benefiting majorly from the new 'searchable' system preferences.
The only thing I'm really looking forward to is the new version of Mail, but it's not something I would spend $140 on -- I'll be getting it free though.
To all those people who suggested Gambas, I knew about it and that's what I meant when I said 'fringe' programs.
Gambas sadly has little to none real support, no Win32 compatibility and it's missing a lot when it comes to widgets/components (think database connectivity and the like).
Sorry, but what competitors to VB6 does the OSS community have?
PythonGTK is nice, but still nowhere near VB6 (look at the complexity of the runtimes). There is a few 'fringe' programs but to be honest you have to get much closer to the bone to do anything on Linux. Even Mono is still lacking true click and drag programming.
I'm not saying this is a bad thing, but there is some areas where Linux still is 'lacking' vs Microsoft/Apple/Sun/whoever, and of course it'll get fixed, but suggesting that people who built CRM suite in '97 would be better off now if they had chose a non-existent Linux/OSS solution is a bit silly.
I'm sure they'd love to have their products run on Linux, but it's quite frankly near impossible for big commercial developers to make anything but high end specialised 3D apps or web apps on Linux.
Something like Photoshop would be an absolute nightmare to port.
Would it be done in Qt, GTK1, GTK2 or raw X widgets? Which printer dialogs would it use -- KDE or GNOME? Which file selectors would it use? How would they keep up, test and fix bugs for GNOME on a 6 month cycle or KDE on a ?? month cycle? How would you have it look nice with the default theme of the desktop?
I can tell you if Adobe ported it the 'slashdoters' would hate it. It would be bloated, slow, buggy and wouldn't fit well into any desktop enviroment. It'd also only be out for x86 and tested on 3 distros max.
The trouble is that at the moment the Linux desktop is moving too fast (with no effort put on old releases of libs or software) at the moment for major software vendors to put out anything but huge 3D apps that are basically their own desktop enviroment, sandboxed from the rest of the system. Personally, I don't think it's a bad thing that Linux is moving really fast, because it's getting closer and closer to Windows or Mac calibre usability with every release, but expecting Adobe to port photoshop, a fairly substainal app with tools that move and break every 6 months is not going to happen.
Slashdot Mirror
And if this car had 100x the sensor and computational capacity of the minimum needed to avert crashing in to walls, then it's ok?
I think we've all done it before, clicked the 'x' button too many times, pressed save on the document you didn't want to keep and don't save on the one you do. It might be user stupidity, but it's avoidable user stupidity.
Right, well just make the 'save copy' save only the last revision.
.tmp files, not just one big-easy-to-backup-and-compress file, which it should do.
Word's autorecovery sucks, it makes hundreds of
The title would be needed, and if they don't, it just means they have to open more documents (and they learn to name it in future).
They wouldn't lose work, which is totally unacceptable apart from a HDD crash these days.
Yes, wonderful.
UK Online is going to be bringing 16mbit/sec uncapped for £30/month sometime in the summer. That's very close to what that is.
Also, BT has 99.6% of the UK population covered with basic DSL, which is the world's highest.
BT as a private company which takes zero funding from the government has rolled broadband out very well.
They are launching a nationwide 8mbit service later this autumn, and trailing ADSL2.
This is going to mean nearly all of the population can have high speed DSL, whereas in france only the big towns can have it. I don't think you'll see villages with less than 300 people in them enabled in france, but you will in the UK.
Yea, but neither have >250k users :).
As for the UK bandwidth thing, bascially you have Bulldog and UK Online who have put their own equipment inside of the exchange, which seriously reduces their costs. Instead of paying BT £45/month line rental for a 2mbit line, they just pay £8/month to rent the basic copper. After they have the DSLAMs and fibre in the exchanges, their ongoing costs are virtually nil.
Also, we have seen BT introduce 'CBC' - capacity based charging, which sees the ISP choosing how much bandwidth they want, and paying for it that way. I think they pay £11/month to BT for the line at the full speed it can go to (currently 2mbit) and then an insane price for their central bandwidth (in 155mbit chunks). This means it's very profitable to cap but not to allow heavy users on.
Do they? Who? BT? Nope. Tiscali? Nope. Wanadoo? Nope. NTL? Nope.
The only major (>250k users) ISP that have an uncapped service are AOL, Blueyonder and Pipex.
Actually, the concept of 'saving' is ridiculous, especially text documents. I saw a computer advertised with a 300GB HDD.
When there is 300GB of space, why on earth isn't the computer saving _every time_ the document is changed, and storing those revisions in a CVS-style way (obviously not with a CVS style interface, perhaps a 'revisions' sidebar with all the major changes listed).
Think about this user who goes and buys a new desktop PC with 300GB of disk space. Assuming he doesn't get infested with adware and they don't install games, they are going to use at most, what, 5GB of space for their documents, pictures and music? 50GB maybe if they also edit video.
There should be no 'save' command in applications any more. There should simply be a 'title document' command and 'save a copy' for when the user needs to copy it to a different computer, floppy disk, whatever.
AOL UK provides a fairly decent no-limits, straight up PPPoA ADSL broadband connection, which you can use with any DSL router or modem -- you don't need their software anymore.
They also provide access to their email via IMAP4.
I wouldn't choose them, but they are extremely well priced in the UK broadband market and have a very good network (in terms of peering, latency and speeds) -- at least on the DSL side of things.
Nearly every other major UK ISP caps users. British Telecom for example has a 512k connection with a 1GB cap for hte same price that AOL does a unlimited one.
Also on the subject of UK Broadband news, UK Online have dropped the price of their 8mbit service to £29.99/month.. which isn't bad at all when you consider it's free activiation and a free 802.11g wireless DSL router.
The reason it will be patented is that if they don't, some other company can just go ahead and patent the same thing, and the only way to rectify it is a long and hard court case which will cost lots and lots of $$$.
Let's assume this company has some old copies of Word 97 or 2000 hanging around that can be reused.
But yes, your point is quite valid otherwise and I should of included it in the comparision.
I disagree entirely.
.exe, .pif etc etc and Firefox means veryl ittle crap is going to come in from the web.
The fact is that the OS is not a big price for a major company. When I can go on dell.com and order 10 2.4GHz Celeron machines with a copy of WinXP Pro for $349, it's not a big deal at all.
Let's say these are for secretarial use. 99% of secreteries know how to use Windows, Word and Outlook.
Let's say I also spend $200-$300 (a day basically) on a techinican to set up a group policy and install Firefox on all these machines. These machines now can't run
Looking at the Windows startup cost it's $349x10 + $300 = $3790.
Now let's see the Linux cost. I'm going to get a maximum of $50 off those Dell machines for chosing Home instead of Pro, I can't 'not have' Windows on it. So that makes it $299/machine. Let's say the cost of installing Linux on each of these is $0.
Now let's look at my army of typists. None of them know how to use Linux/GNOME, OpenOffice or Evolution. So I train them. I hire a training guy to come in for a day to give them a crash course on how to use Linux, and he charges me $200. However, I've got to pay my typists anyway, $100/ea for the day. So that's $100x10 + $200 = $1200.
Linux startup cost: $2990 + $1200 = $4190. Windows wins.
Now, this is probably a bad example, but training costs, which are not going to change for the short to medium term, are very expensive.
For many small businesses this is the situation they have, and it's even worse if they have specialist apps they need to run on Windows.
So saying 'Linux has TCO ownage on Windows' is a bit unfair. It's very much true (IMO) for servers and workstations. But for average 'business desktops' I don't think it is.
I'd love to agree with you, but sadly graphics and physics engines are huge money-makers.
The Doom3 engine over its lifetime will sell for many times more than the Doom3 game itself sold for. I'd guess Valve wants to try and do this also with Source.
Not to mention both of these games I belive use the Havok physics engine which is done by a completely different company - I think all Havok does _is_ physics engines. Suggesting they GPL their only money maker is not going to work sadly.
I think Open Source's best place is going to be the building blocks for the apps that run on top of them. For example, a web application which you deploy via Firefox running on Linux. Or a database app written on top of pgSQL or mySQL. This is going to be a two way process - the people developing the 'front ends' to these blocks are going to be improving the back-end - it's in their interest.
Look at Google though.
It must of took, what, 3 years at least to even start making money with them. Now they have a market cap of over $40billion.
But yes, your advise is sound, but don't forget some of the best companies have been a complete gamble. Keeping costs low is always very important and if you can save $800 on your server setup with OSS, that could be the extra week than you need to survive in business and get that big contract.
I agree -- Python is fantastic for quickly building small apps, or even much larger ones.
The problem arises in Python's web programming support. The documentation is pretty much non-existent and you can soon get module-overload when you are importing more and more modules to do fairly simple stuff in web apps.
Sometimes I just think while Python is most certainly a far better designed language, PHP/ASP.NET (C#) seems much more 'pratical', and it's definitely much easier to quickly build web apps in.
Is there much effort to improve Python's web support? A manual with similar completion of php.net would help it go a lot lot further.
I think they mean more selling chip layouts to other companies for manufacture.
This is going to be a huge industry. If they could produce, say, a 800MHz CPU which ran on 1W or 0.5W of power and had sensible float performance, it could easily sell exceedingly well.
There must be millions upon millions of devices that require more than just PIC-level performance but low power consumption. Things like digital TV decoders -- the video itself can be decoded with a seperate chip but the amount of interactivity that will be delivered in the future is going to be immense.
Ok. I take what you say onboard. I do like what you say about automator and spotlight, and to be honest maybe I was not giving enough thought to them (especially automator, from Apple's screenshots it looks to me to be little more than some terribly limited graphical applescript).
Spotlight sounds intresting. I have used somewhat similar apps like Quicksilver (and some other one I can't think of) and it promised the same and failed horribly. Inaccurate, slow and unresponsive. So,
I really think that Apple is pushing the hardware _too_ hard. I think it will be amazing for someone running it on a dual core Powermac G5 3.0GHz. But for the millions (the majority) who are running on G3 iBooks or G4 iMacs/Powerbooks/eMacs/iBooks then I think it's going to be a pretty bad experience.
Sure, it sounds very 'science-fiction cool' to have the system 'automatically' choose the best system and reoptimise it for it, but it's really not very practical.
The fact is that the GPU will be 10-25x+ faster than Altivec vector processing unit. Altivec will probably be 2-5x faster than the CPU. If this is going to mean you get 5fps on an effect vs 100fps or more it's going to be a big deal. If you can only get 5fps with the vast majority of the userbase you are not going to spend loads of extra programming time and money creating a completetly different codepath for pixel shaders vs altivec vs normal, and scaling down the effects as you go along.
The thing is that Apple could put a NVidia 5200 on it and it would be able to do this. But instead they continue shipping the frankly bollocks Radeon 9200. How much would it cost to upgrade it? $10, $15?
Yes, but with the G4 the FSB is choked to hell already. Chucking down really complex pixel shading effects to do on the _CPU_ is going to slow it down so much it's going to be unusable for anything more than the most basic effects. If it was easy and fast to do on the CPU we would of already seen it. The fact is that you need a dedicated unit for this, on the GPU.
I agree entirely on your second point, apart from the time frame. I'd guess it'll be 3 years before the majority of the Mac userbase has a pixel shader onboard. I mean, after 5 years of OSX only 60-80% of the userbase has moved, and all that takes is a CD inserted into the drive, and not a new Mac that many will need for PS support.
I'm sure some of the new features will be useful, but to me it seems that Tiger is being completely overhyped.
Wow.
Sorry, but for most people CoreImage and CoreVideo is going to be utterly useless. Apple still ships shit, shit, shit video processors on the iBook, Mac Mini and only the latest generation Powerbooks, PMs and iMac have the much-needed Pixel Shader on their GPUs. I'd guess probably 10-20% of the Mac userbase uses a Powerbook latest revision, PM G5, or iMac G5. The iBook was Apple's best selling Mac a few months back and I'm sure that the Mac Mini will replace it.
So are you honestly going to tell me developers are going to bother developing with features that only 10-20% of their already small userbase can use?
Personally I don't see any one feature that Tiger has that I really want. Hopefully it'll be a lot more polished and have some nice performance increases, but the vast majoirty of stuff in Tiger is totally useless to me: I don't need spotlight since I organise my stuff well, I don't use Safari for anything more than basic browsing (I have a perfectly good RSS client already, thanks), I won't be using automator, quicktime or benefiting majorly from the new 'searchable' system preferences.
The only thing I'm really looking forward to is the new version of Mail, but it's not something I would spend $140 on -- I'll be getting it free though.
To all those people who suggested Gambas, I knew about it and that's what I meant when I said 'fringe' programs.
Gambas sadly has little to none real support, no Win32 compatibility and it's missing a lot when it comes to widgets/components (think database connectivity and the like).
It does look promising though...
Acrobat reader is a little bit different to Photoshop. I wouldn't be suprised if Photoshop is 10 or even 25 times as large as reader.
Sorry, but what competitors to VB6 does the OSS community have?
PythonGTK is nice, but still nowhere near VB6 (look at the complexity of the runtimes). There is a few 'fringe' programs but to be honest you have to get much closer to the bone to do anything on Linux. Even Mono is still lacking true click and drag programming.
I'm not saying this is a bad thing, but there is some areas where Linux still is 'lacking' vs Microsoft/Apple/Sun/whoever, and of course it'll get fixed, but suggesting that people who built CRM suite in '97 would be better off now if they had chose a non-existent Linux/OSS solution is a bit silly.
I'm sure they'd love to have their products run on Linux, but it's quite frankly near impossible for big commercial developers to make anything but high end specialised 3D apps or web apps on Linux.
Something like Photoshop would be an absolute nightmare to port.
Would it be done in Qt, GTK1, GTK2 or raw X widgets? Which printer dialogs would it use -- KDE or GNOME? Which file selectors would it use? How would they keep up, test and fix bugs for GNOME on a 6 month cycle or KDE on a ?? month cycle? How would you have it look nice with the default theme of the desktop?
I can tell you if Adobe ported it the 'slashdoters' would hate it. It would be bloated, slow, buggy and wouldn't fit well into any desktop enviroment. It'd also only be out for x86 and tested on 3 distros max.
The trouble is that at the moment the Linux desktop is moving too fast (with no effort put on old releases of libs or software) at the moment for major software vendors to put out anything but huge 3D apps that are basically their own desktop enviroment, sandboxed from the rest of the system. Personally, I don't think it's a bad thing that Linux is moving really fast, because it's getting closer and closer to Windows or Mac calibre usability with every release, but expecting Adobe to port photoshop, a fairly substainal app with tools that move and break every 6 months is not going to happen.
I was using market cap, not share price.
I know it's not a perfect valuation of the company but for publically listed ones it's really the only one...