Well, sourcesafe stores everything in a proprietary database format, so a little bit of db corruption can easily eliminate "50,000 lines of legacy code". Unintentionally.
Wrong. Javascript is supported server-side by MS IIS, Sun One (was chilisoft) ASP and stryon (was halcyon) iASP - wiht the latter two running on a wide range of web servers.
Write your javascript well and you can use the same functions/objects (eg. form validators) on the client and the server.
Opera (at least up to v6) doesn't support DOM except for reading, which means it is pretty much useless for DHTML - the Dynamic bit is kind of lost if you can't actually change anything.
GPL is very restrictive in what it can be linked to. Some people may have code under other licences (eg. old-style BSD with advertising clause) that they want to combine with the new released code. If it is GPL only, then they can't do that, so using GPL is therefore discriminating against some developers and not benefiting everyone equally.
Eg.: if a reference implementation for a new network protocol were released under the GPL, it could be implemented in Linux very quickly, but FreeBSD would have to clean-room rewrite it. I don't think that is at all fair - even though I use Linux in preference to *BSD.
LGPL would be better from this point of view, but still not ideal. Best is public domain or very minimal BSD style licence which can be combined with almost any other code (including GPL) without legal problem.
Yes, that would include proprietary code, but to exclude it without excluding some (free/open-source) developers from the benefits, you would have to write some form of new licence that is legally compatible with all other open-source/free licences without being compatible with any proprietary licences - something I reckon is impossible.
You have indeed made a mistake, but (judging by other replies) it is a common one.
To quote ESR's paper:
I believed that the most important software (operating systems and really large tools like the Emacs programming editor) needed to be built like cathedrals, carefully crafted by individual wizards or small bands of mages working in splendid isolation, with no beta to be released before its time.
I don't see how this can possibly be interpreted as Proprietary == Cathedral. There are plenty of other hints that it's Linux == Bazaar, GNU == Cathedral.
Probably most proprietary software is also cathedral, simply due to small number of developers with eyes on the code, but there are organisations which have large enough development teams that they could possibly run a project bazaar-style without opening it up outside the organisation.
No, the aim should be to make the web accessible to people using standard browsers, not non-standard. The standards already exist, and so do browsers that support them - web sites written for non-standard browsers are the problem not the solution.
Delivery might not cost any more, but for most web sites the cost of development dwarfs the cost of delivery.
The costs of converting existing sites will clearly not be zero. Even for new sites designed from the start with accessibility in mind you may well have to retrain existing designers, buy new design tools, new testing tools, add new tests to your approvals process. All costs.
Point is that even if your mobile is off, if you _could_ get a signal then there is a tower transmitting at you. Those transmissions bounce off you whether you have a phone or not.
Using handset signals for this is probably a non starter as you don't know the handset position accurately (to calculate your reflections) and the signal strength is much lower.
Nope this is different - read the article. It is not about tracking phones (which they already do) it's about tracking objects using the phone signals.
It's basically radar but using cellphone transmissions as the source signal, so you don't need to put up radar transmitters everywhere because the telcos have done it for you.
Nope this is different - read the article. It's basically radar but using cellphone transmissions as the source signal, so you don't need to put up radar transmitters everywhere because the telcos have done it for you.
HKLM\SYSTEM\CurrentControlSet\Control\ContentInd ex \FilterFilesWithUnknownExtensions
to 1 (maybe you have this set ?) then all "unknown" file types get searched as text, and on my system this searches.java files where it doesn't if it's set to 0.
Do you by any chance _not_ have jbuilder installed (I do) ? I reckon JBuilder may have messed up the.java registry keys somehow.
Tried this thinking you must be wrong, but nope looks like you're right.
It doesn't appear to look inside.java files at all - rename the file to.java.txt and it finds it straight away. For some reason it is treating.java as binary rather than text. Must be some registry key you can change to control that.
FWIW other MS apps (eg. visual studio) find text in.java files just fine.
The whole point of the articles is that KDE League folks != KDE (actual developer doing the work type) folks.
So it's not "half-dozen lazy Europeans who know C++" running the corp, it's some other (mostly undisclosed) US folks (residency normally a requirement for at least some directors?) running it and taking the money, while the "lazy Europeans" continue to produce the goods without the benefit of the cash.
My bad - I'm thinking of NTbackup on 2000. Been a long while (in IT time) since I used it on NT4 and I can't remember what that version did. In some ways 2000 is worse as exchange becomes tightly integrated with active directory and getting email back after you've lost AD is horror story even if you have a workign exchange backup.
I know about loss of single-instance storage with psts, but after several days of crashed drives, mysteriously unreadable backups, broken tape drives, etc. it is a great relief to just give a pst to each user and say "there's your old mail". that's with tens of users though, obviously with hundreds or more it may be a different story.
NTBackup can use files (including network) as target media.
Not sure i trust it to backup exchange correctly (ie. with something that actually restores) though.
For exchange backup you can also either dimount mailstore and copy db and transaction log files or use the exmerge tool (can run in batch mode from scripts or scheduler) to export mailboxes to.pst files which can then be backed up like normal files.
Probably wouldn't work because a fixed/hidden camera doesn't select what is photographed so there is no intent to film underage.
Running round sticking cameras under young girls skirts would imply intent and so would get prosecuted.
On the other hand, if it did work the way you suggest, it would be a nice fight back against cctv everywhere - simply get a 17y/o to flash tits (face to identify, tits=porn) at all the cctv surviellance cameras you can find, all cctv operators are now kiddyporn possesors and cctv effectively becomes illegal...
better still, build the psu into ups, always bugs me that the ups sits there converting it's output to AC and conditioning it, when it then gets turned straight back into low volts DC at the PSU at the other end of the power cord. Most modern home/office peripherals, and with a TFT even the monitor, also want DC.
of course you don't want a combined ups/psu for carrying to a lan party with cff case - weight wise they aint easily portable:-)
Most commercial aircraft currently flying were designed decades before laptops and mobile phones. Almost every elecrical system will require testing/redesign, and there are literally millions of them on an large aircraft, and literally miles of wiring. Then you'd have to re-certify the modified design as safe for flight, and that's before you even get to the cost to refit each aircraft, which is probably going to be a significant fraction of the cost of the aircraft.
And of course the airline industry has loads of spare cash at the moment doesn't it... ummm.
Millitary aircraft are likely to be more resistant, but mostly they are still old designs that wouldn't have considered the specific characteristics of laptop/mobile rf from within the aircraft. That said, I reckon I'd be a lot more prepared to use a mobile in a figther jet simply on the basis that you're sat on an ejector seat that definitely doesn't use any fancy fly-by-wire or rf electronics.
On cruise missiles you have to remember that your laptop / mobile is inside the metal tube of the plane, along with the sensitive electronics - if the bad guys could fly up to the missile, open it up, and stick a laptop and cellphone inside it, it probably would confuse the guidance - but if you can get that close to it there are easier ways to knock it out of the sky.
Erm, I think the catholics have been after Paisley for years. He's porbably only still around because they figure his ranting does the unionists more harm than good.
They also move up/down as well as back - provided you get the higher spec models. For tall drivers this is absolutely essential. Also the steering wheel should have height/reach adjust to get it away from your knees.
I'm (only!) 6'2, but I have quite reasonable clearance above-head and knees-steering wheel in our (2001) golf. I normally drive it with seat right down and right back, but I can cope (ie. its not as comfortable but my legs/head are not actually jammed against stuff) with it slightly raised and slightly forward, if we need to get stuff in the back.
Sorry, you are wrong, IE5 was wrong and Mozilla (and IE5 mac and IE6 and Opera) are right.
A 32x32 element with 16 pixel border should in fact be 32x32 content _plus_ 16 pixel border.
Does this make it a right pain to use percentage lengths ? - yes but it is what the standard says. See eg. http://www.alistapart.com/stories/journey/4.html In my experience things are now much better than they were, with reasonable care not to use IE specific sutff (and the MS docs do tell you whether each element / attribute is standard or not) you can develop on IE and have it work >90% correct first time on Mozilla or other Gecko-based browsers.
Opera is way behind the other two though (often Konqueror does a better job) I refuse to do dynamic pages using a reload and loads of document.write() when DOM1 has been a standard since 1998. Also a lot of css (eg. overflow, display) which you might use to get rid of frames & tables doesn't work in Opera.
Re:Found a Huge Hole and a Contradiction
on
GPL FAQ
·
· Score: 1
Yes, but if I had written it so that it uses a standard Win32API or COM object that is normally shipped with Windows and your component replaces that, at what fault am I?
My component does not replace the API, it is a plugin which extends it. Since the API is designed to be extended in this way, by using it you are effectively linking your program to arbitrary libraries. Now, according the FSF interpretation of the GPL, the GPL does not allow you to do this. the only logical conclusion is that you cannot therefore use such an API in a GPLed program... which pretty much means you cannot distribute a GPLed windows program.
You (the user) can't make me (the developer) break the GPL.
The FSF would say that you, the developer, cannot hide behind the actions of the user, and that if you write code which links GPL & proprietary code when the user runs it, then you are in violation.
Re:Things RMS didn't forsee in 1984
on
GPL FAQ
·
· Score: 1
DLLs and object brokers perhaps not, but there were a lot of non-unix systems where linking was significantly different (eg. think TSRs and overlays on DOS, or cooperative-multitasking on Macs).
I think the main problem is that RMS wrote the GPL for the GNU system which was to be
1980s-unix-like, and it makes sense for such systems, but taken outside of them it starts to become unclear.
It doesn't help matters that even today when issuing information like the recent FAQ, the FSF still explains the GPL only in the context of (80s) unix systems. Clarifying what constitutes a separate program in terms of "fork/exec" or separate memory spaces is no use on systems which don't have fork/exec or per-process memory spaces, even before we get into dynamic linking, location-transparency etc.
In the end the courts may have to decide what is or is not a separate program in the context of the GPL on each operating system. The way Microsoft describe and implement their licensing for dlls/COM components indicates that they view them as separate entities (not derivative works) wrt. copyright, and, since they wrote the OS, that view would seem to me to be rather important when deciding what constitutes a "program" on that OS (I believe it is a Windows dll that is at issue in this case).
Re:Found a Huge Hole and a Contradiction
on
GPL FAQ
·
· Score: 1
the API is probably part of the Win32 API set, which is distributed in every copy of Windows 9x/NT/2000. Since it's part of the operating system the source doesn't have to be released.
However, almost all windows programs use Windows APIs for which the implementation does not ship with the OS (or at least, not always). There is no libc provided as standard with Windows - you have to ship it with your executable, but the GPL says you can use the OS exception unless that component itself accompanies the executable, so you can link to it but you have to rely on your user getting it from somewhere else in order to make your.exe run. At least that is what I think it says.
Second problem is that the Win32 API is itself extensively based on dynamic linking (via COM) and API calls may result in third party installed dlls being dynamically linked into your executable. This includes dlls you (the author) have never seen or used, that may not even have existed when you wrote your program - this is what dynamic linking means.
So, I can produce a (non-GPL) dll which is linked to your Windows GPL program whenever the program makes a system API call on a machine with my dll installed. So, your GPL program now uses propritary library 'FOO' that provides some functionality that is not part of the o/s or compiler... but that (as you say) is not allowed...
Slashdot Mirror
Well, sourcesafe stores everything in a proprietary database format, so a little bit of db corruption can easily eliminate "50,000 lines of legacy code". Unintentionally.
Wrong. Javascript is supported server-side by MS IIS, Sun One (was chilisoft) ASP and stryon (was halcyon) iASP - wiht the latter two running on a wide range of web servers.
Write your javascript well and you can use the same functions/objects (eg. form validators) on the client and the server.
Opera (at least up to v6) doesn't support DOM except for reading, which means it is pretty much useless for DHTML - the Dynamic bit is kind of lost if you can't actually change anything.
which is really odd seeing as mostly on /. people don't read the article before commenting...
GPL is very restrictive in what it can be linked to. Some people may have code under other licences (eg. old-style BSD with advertising clause) that they want to combine with the new released code. If it is GPL only, then they can't do that, so using GPL is therefore discriminating against some developers and not benefiting everyone equally.
Eg.: if a reference implementation for a new network protocol were released under the GPL, it could be implemented in Linux very quickly, but FreeBSD would have to clean-room rewrite it. I don't think that is at all fair - even though I use Linux in preference to *BSD.
LGPL would be better from this point of view, but still not ideal. Best is public domain or very minimal BSD style licence which can be combined with almost any other code (including GPL) without legal problem.
Yes, that would include proprietary code, but to exclude it without excluding some (free/open-source) developers from the benefits, you would have to write some form of new licence that is legally compatible with all other open-source/free licences without being compatible with any proprietary licences - something I reckon is impossible.
To quote ESR's paper:I don't see how this can possibly be interpreted as Proprietary == Cathedral. There are plenty of other hints that it's Linux == Bazaar, GNU == Cathedral.
Probably most proprietary software is also cathedral, simply due to small number of developers with eyes on the code, but there are organisations which have large enough development teams that they could possibly run a project bazaar-style without opening it up outside the organisation.
No, the aim should be to make the web accessible to people using standard browsers, not non-standard. The standards already exist, and so do browsers that support them - web sites written for non-standard browsers are the problem not the solution.
Delivery might not cost any more, but for most web sites the cost of development dwarfs the cost of delivery.
The costs of converting existing sites will clearly not be zero. Even for new sites designed from the start with accessibility in mind you may well have to retrain existing designers, buy new design tools, new testing tools, add new tests to your approvals process. All costs.
Point is that even if your mobile is off, if you _could_ get a signal then there is a tower transmitting at you. Those transmissions bounce off you whether you have a phone or not.
Using handset signals for this is probably a non starter as you don't know the handset position accurately (to calculate your reflections) and the signal strength is much lower.
Nope this is different - read the article. It is not about tracking phones (which they already do) it's about tracking objects using the phone signals.
It's basically radar but using cellphone transmissions as the source signal, so you don't need to put up radar transmitters everywhere because the telcos have done it for you.
Nope this is different - read the article. It's basically radar but using cellphone transmissions as the source signal, so you don't need to put up radar transmitters everywhere because the telcos have done it for you.
If you set
d ex \FilterFilesWithUnknownExtensions
.java files where it doesn't if it's set to 0.
.java registry keys somehow.
HKLM\SYSTEM\CurrentControlSet\Control\ContentIn
to 1 (maybe you have this set ?) then all "unknown" file types get searched as text, and on my system this searches
Do you by any chance _not_ have jbuilder installed (I do) ? I reckon JBuilder may have messed up the
Tried this thinking you must be wrong, but nope looks like you're right.
.java files at all - rename the file to .java.txt and it finds it straight away. For some reason it is treating .java as binary rather than text. Must be some registry key you can change to control that.
.java files just fine.
It doesn't appear to look inside
FWIW other MS apps (eg. visual studio) find text in
Why reply to a troll, ah what the hell:
The whole point of the articles is that KDE League folks != KDE (actual developer doing the work type) folks.
So it's not "half-dozen lazy Europeans who know C++" running the corp, it's some other (mostly undisclosed) US folks (residency normally a requirement for at least some directors?) running it and taking the money, while the "lazy Europeans" continue to produce the goods without the benefit of the cash.
My bad - I'm thinking of NTbackup on 2000. Been a long while (in IT time) since I used it on NT4 and I can't remember what that version did. In some ways 2000 is worse as exchange becomes tightly integrated with active directory and getting email back after you've lost AD is horror story even if you have a workign exchange backup.
I know about loss of single-instance storage with psts, but after several days of crashed drives, mysteriously unreadable backups, broken tape drives, etc. it is a great relief to just give a pst to each user and say "there's your old mail". that's with tens of users though, obviously with hundreds or more it may be a different story.
NTBackup can use files (including network) as target media.
.pst files which can then be backed up like normal files.
Not sure i trust it to backup exchange correctly (ie. with something that actually restores) though.
For exchange backup you can also either dimount mailstore and copy db and transaction log files or use the exmerge tool (can run in batch mode from scripts or scheduler) to export mailboxes to
Probably wouldn't work because a fixed/hidden camera doesn't select what is photographed so there is no intent to film underage.
Running round sticking cameras under young girls skirts would imply intent and so would get prosecuted.
On the other hand, if it did work the way you suggest, it would be a nice fight back against cctv everywhere - simply get a 17y/o to flash tits (face to identify, tits=porn) at all the cctv surviellance cameras you can find, all cctv operators are now kiddyporn possesors and cctv effectively becomes illegal...
better still, build the psu into ups, always bugs me that the ups sits there converting it's output to AC and conditioning it, when it then gets turned straight back into low volts DC at the PSU at the other end of the power cord. Most modern home/office peripherals, and with a TFT even the monitor, also want DC.
:-)
of course you don't want a combined ups/psu for carrying to a lan party with cff case - weight wise they aint easily portable
Most commercial aircraft currently flying were designed decades before laptops and mobile phones. Almost every elecrical system will require testing/redesign, and there are literally millions of them on an large aircraft, and literally miles of wiring. Then you'd have to re-certify the modified design as safe for flight, and that's before you even get to the cost to refit each aircraft, which is probably going to be a significant fraction of the cost of the aircraft.
And of course the airline industry has loads of spare cash at the moment doesn't it... ummm.
Millitary aircraft are likely to be more resistant, but mostly they are still old designs that wouldn't have considered the specific characteristics of laptop/mobile rf from within the aircraft. That said, I reckon I'd be a lot more prepared to use a mobile in a figther jet simply on the basis that you're sat on an ejector seat that definitely doesn't use any fancy fly-by-wire or rf electronics.
On cruise missiles you have to remember that your laptop / mobile is inside the metal tube of the plane, along with the sensitive electronics - if the bad guys could fly up to the missile, open it up, and stick a laptop and cellphone inside it, it probably would confuse the guidance - but if you can get that close to it there are easier ways to knock it out of the sky.
Erm, I think the catholics have been after Paisley for years. He's porbably only still around because they figure his ranting does the unionists more harm than good.
They also move up/down as well as back - provided you get the higher spec models. For tall drivers this is absolutely essential. Also the steering wheel should have height/reach adjust to get it away from your knees.
I'm (only!) 6'2, but I have quite reasonable clearance above-head and knees-steering wheel in our (2001) golf. I normally drive it with seat right down and right back, but I can cope (ie. its not as comfortable but my legs/head are not actually jammed against stuff) with it slightly raised and slightly forward, if we need to get stuff in the back.
Sorry, you are wrong, IE5 was wrong and Mozilla (and IE5 mac and IE6 and Opera) are right.
A 32x32 element with 16 pixel border should in fact be 32x32 content _plus_ 16 pixel border.
Does this make it a right pain to use percentage lengths ? - yes but it is what the standard says.
See eg. http://www.alistapart.com/stories/journey/4.html
In my experience things are now much better than they were, with reasonable care not to use IE specific sutff (and the MS docs do tell you whether each element / attribute is standard or not) you can develop on IE and have it work >90% correct first time on Mozilla or other Gecko-based browsers.
Opera is way behind the other two though (often Konqueror does a better job) I refuse to do dynamic pages using a reload and loads of document.write() when DOM1 has been a standard since 1998. Also a lot of css (eg. overflow, display) which you might use to get rid of frames & tables doesn't work in Opera.
Yes, but if I had written it so that it uses a standard Win32API or COM object that is normally shipped with Windows and your component replaces that, at what fault am I?
My component does not replace the API, it is a plugin which extends it. Since the API is designed to be extended in this way, by using it you are effectively linking your program to arbitrary libraries. Now, according the FSF interpretation of the GPL, the GPL does not allow you to do this. the only logical conclusion is that you cannot therefore use such an API in a GPLed program... which pretty much means you cannot distribute a GPLed windows program.
You (the user) can't make me (the developer) break the GPL.
The FSF would say that you, the developer, cannot hide behind the actions of the user, and that if you write code which links GPL & proprietary code when the user runs it, then you are in violation.
DLLs and object brokers perhaps not, but there were a lot of non-unix systems where linking was significantly different (eg. think TSRs and overlays on DOS, or cooperative-multitasking on Macs).
I think the main problem is that RMS wrote the GPL for the GNU system which was to be 1980s-unix-like, and it makes sense for such systems, but taken outside of them it starts to become unclear.
It doesn't help matters that even today when issuing information like the recent FAQ, the FSF still explains the GPL only in the context of (80s) unix systems. Clarifying what constitutes a separate program in terms of "fork/exec" or separate memory spaces is no use on systems which don't have fork/exec or per-process memory spaces, even before we get into dynamic linking, location-transparency etc.
In the end the courts may have to decide what is or is not a separate program in the context of the GPL on each operating system. The way Microsoft describe and implement their licensing for dlls/COM components indicates that they view them as separate entities (not derivative works) wrt. copyright, and, since they wrote the OS, that view would seem to me to be rather important when deciding what constitutes a "program" on that OS (I believe it is a Windows dll that is at issue in this case).
the API is probably part of the Win32 API set, which is distributed in every copy of Windows 9x/NT/2000. Since it's part of the operating system the source doesn't have to be released.
.exe run. At least that is what I think it says.
However, almost all windows programs use Windows APIs for which the implementation does not ship with the OS (or at least, not always). There is no libc provided as standard with Windows - you have to ship it with your executable, but the GPL says you can use the OS exception unless that component itself accompanies the executable, so you can link to it but you have to rely on your user getting it from somewhere else in order to make your
Second problem is that the Win32 API is itself extensively based on dynamic linking (via COM) and API calls may result in third party installed dlls being dynamically linked into your executable. This includes dlls you (the author) have never seen or used, that may not even have existed when you wrote your program - this is what dynamic linking means.
So, I can produce a (non-GPL) dll which is linked to your Windows GPL program whenever the program makes a system API call on a machine with my dll installed. So, your GPL program now uses propritary library 'FOO' that provides some functionality that is not part of the o/s or compiler... but that (as you say) is not allowed...