> 1. The license only restricts your ability to take java, change the name and call it your product, then start charging for it, without paying license and royalty fees to Sun.
Funny I could swear I cut and pasted this from the SUN site.
"Modified source code cannot be distributed without the express written permission of Sun Microsystems, Inc."
So basically you can read the code to find out how it works, but you can't distribute bug fixes, or enhancements, you can't port useful bits of the code to other Java implementations or other software (indeed writing your own version after you've read the source code might be legally risky I suspect). It makes no mention of whether you charge for it.
Indeed the licence attempts to risk redistribution of modified binaries "internally", so modifying the code for your companies own private use may be a licence infringement.
"Here is the source code, look and admire, but don't touch it."
Compare and contrast the licence with the other 16,000 odd packages in Debian.
Dell's personal worth 1.6 x 10^10 Investment in Redhat 10^8
What proportion did you say again;)
Don't get me wrong I think your comment is right to the point. But this would be like me getting a mortage of 100,000USD, and then investing 600USD in Redhat.
Because of the numbers involved it sounds big finance, but without knowing the portfolio you can't know if it is hedging, or represents a major personal commitment to Redhat.
Michael Dell is a smart business man, I'm betting he sees it as a long term bet. If you buy the "GNU/Linux will be the next big thing" (and it already is in the server market), and look for a company to invest in to gain from this, Redhat is the obvious choice (IBM is just too big, they could succeed in GNU/Linux big time and only affect the business value by a few points, many of the little exciting companies are just too unpredictable when you have 100 million to invest).
Virus scanning is so obviously not the fix to this problem I'm not sure where to start.
I don't run virus checking on this PC because there are no major viruses in the wild for the operating system, because people deal with the issues properly in a timely fashion when they arise.
Sure it is vulnerable to a new types of attack in a new vulnberability, but virus checking my email is very unlikely to stop that.
Virus checking only protects people who don't deal with issues in a timely fashion, or whose vendor doesn't.
Virus checking for email is basically a short term fudge till people get decent email clients, or update the existing one.
ISPs could refuse to send email for people using old versions of Outlook/Outlook Express. This would be just as effective in stopping email bourne viruses I suspect, and far easier to implement.
The spam problem viruses tend to set up SMTP proxies that allow anonymous amplification of the injected message. Not a true SMTP server, but not really a conventional client.
"That is properly the smartes idea in this thread."
Till the spammer send out an Official looking Microsoft CD that owns the recipients PC.
Microsoft will send patch CDs out to anyone that asks for them AFAIK, which is the right way to do it. Most good computer magazines also often ship copies of Microsoft patches on cover CDs/DVDs.
At the end of the day the end user still has to take a positive action till SP2 appeared, even if it is "insert CD", which alas it often isn't.
Each of these CD ought to trigger a patch level test and provide a simple "You are out of date?" "Get Upto Date" "I really know what I'm doing" dialog. Because the problem users are mostly the ones who would rarely click on a "I know what I'm doing" button.
Yes but here as pointed out this is not a big problem in terms of numbers.
You see perhaps one or two boxes attempting a brute force ssh password attack a day on a well hosted server.
Sure it creates a lot of log file mess, but that is because by default it logs every attempt, and they often try 200 root passwords in one go.
Dshield data suggests they see 600 sources looking for port 22 a day, if we assume a fair number of these are just port scans, it doesn't leave much activity, and I've seen it suggested this is only semiautomated.
Definitely the worst current problem for GNU/Linux. But fairly stable, so presumably this is mostly people who don't give a monkeys that their box has been owned.
One day someone will get pissed off and try to brute force the passwords on these attackers;)
I think some GNU/Linux users are complacent to about security, but I doubt we'll ever see anything on the scale of the Sober viruses till there is one big sucessful Desktop distribution.
I just use Postfix MIME type header inspection to reject Windows executable content types.
It is literally a three line change to Postfix, one to say use this file of REGEXP, and two regular expressions. Linked from the postfix.org site.
Do that in advanced of the antivirus, and you'll reduce the load to just unpacking those damn "zip" file attachments.
For my own email I reject ZIP file attachments as well. Realised I was getting many thousands of viruses to every genuine ZIP file. And most stuff that matters was a.tar.gz extension anyway;)
"It's interesting because it means that there are still enough unpatched machines out there for a worm to gain serious traction without uncovering new technical vulnerabilities."
You'd have to be living on another planet for this to be news.
Depressingly local computer shops were selling Windows with only SP1 installed a long time after SP2 was release and certainly within the past 6 months.
The answer to "why now" is on the Veritest web page.
The paper is hilarious if you actually read it.
Key points...
They looked for tips on optimising Apache on the Redhat website ?!(guys next time try httpd.apache.org).
They found that Apache 2 was 50% faster than Apache 1 (without any tuning, Apache 2 offers a selection of threading models so a fair comparion would have tried each in turn, with or without tuning), so presumably didn't do any further tests with that in case it made MS look bad.
They tuned Windows for the server but effectively plead ignorance of how to tune Redhat for the server.
My guess is even then, on this hardware, sensible tuning of the kind they did to Windows would have made Apache comparable or better.
There were issues with this hardware selection at the time. Driving Gigabit ethernet is pretty demanding stuff, and you need drivers that can handle interrupt load.
However anyone who actually needs an 8 CPU machine to serve gigabit websites will probably do their own benchmarking and tuning, in the unlikely event their application software gives them a free choice of platform.
When my employers online business is big enough to need gigabit hosting, we'll probably still have SQUID on Linux with shed loads of memory accelerating the static content, because hardware is cheaper than rewriting the corporate applications.
As someone using Unix for nearly 20 years, I'd have to ask also "is the change worth it?".
Whilst there appears to be some gain in consistency in going the launchd way, there is no particular gain in functionality or flexibility over the a combination of initd, and xinetd, and the SVR4(ish) init scripts. Indeed scripts can do everything, and I'd be surprised if some of the things Apple calls aren't scripts precisely to do "extra stuff".
Others have proposed different daemon handling schemes before, such as DJB. Each scheme has its merits, and it is always cleaner to start again.
But it is easy to underestimate the amount of work in moving a big GNU/Linux distro to a different start-up system, not just scripts but documentation as well, and the loss of "know how" as all those thousands of usenet posts no longer answer the question of why o and so application doesn't start on boot.
That said I've never been a fan of the bastardised Linux way of doing start-up scripts. I remember when run levels were really levels.
Inefficiencies of lauching a shell as the slow bit?
Well it is possibly non-trivial I see about 1 sec for launching 100 shells sequentially on my desktop, but I'd always been led to believe the predominant reason for slow deskstop start-up on GNU/Linux was the sheer number of files required to get a working desktop together (~500), which could only be reduced by eliminating seek time (boot from flash), or a radical restructuring of XFree86.
Certainly the last time I saw MACOS X boot it didn't look especially fast.
I vow to always reply to these threads when I see them with my latest stats, just to hack people off;)
Yesterday 81.1% of hits on our site were IE. Firefox in second 11.1%.
Safari, "Other Gecko", and the ubiquitous "Unknown" were the only other browser grouping to score more than 1%.
Konqueror barely exceeded wget 458 versus 312, both show as 0.0% when rounded to one decimal place by our stats program.
The OS stats continue to suggest a monopoly position for Microsoft, although I need to dig into that "3.1% unknown" which is the largest single alternative, to see how "Unknown" it really is, and how much I need to update the patterns.
Saying KDE is not cheating, we are talking about Window Manager levels, it would be the appropriate tool to compare to an Apple Desktop.
Or is AppleScript going to magically allow control of all apps written in any combination of languages and tools that can be compiled on Macos X?
But then we are buried beneath a thread where someone seemed to suggest Linux in response to "scripting". Scripting predates me, and I predate Linux somewhat. I've only briefly used Multics in anger, but it felt terribly familiar.
Be assured whatever Apple release it will make "Microsoft Test" look rather feeble;)
One is purely performance/hardware, it has long been possible to do a free text search of all the contents of your computer (security issues aside - this is very hard to do securely in a multiuser environment if anyone cares about security/privacy), just no one could be bothered to sacrifice the amount of disk space, and CPU to do this historically. Thus this is purely a technology whose time has come, previously Microsoft and others were mostly indexing metadata (optionally, like you'd bother to switch it off these days, but 5 years ago you may well have chosen to), or select portions of the system.
When Google announced they were doing this I downloaded a free software product that already did this for GNU/Linux for comparison, and yes it worked, it also took about a day to index my system as root, doubled my disk space usage, and needed to refresh its indexes (which it chose to do overnight), and made everyones content visible (if only indirectly) to anyone who could query the database.
As such indexing is a natural progression of computing, but as any database person will tell you indexing has a big resource cost. These days no one cares if their 150GB hard disk is 20% used instead of 10%, and if writing a file takes 10000 operations instead of 100.
The other is integration, Apple as always have a truely innovative integration between search and desktop. I don't always buy their "usability" features (I like menus in the Window they relate to, saves a LOT of mouse mileage). But I think they probably have the edge here, and will retain it if only because of the relatively small application base installed on most Apple desktops.
Problem is it isn't clear whose bad programming. This slow switch only happens on Windows, I suspect as OS/Library bug, but it is possible it is something in the Windows version.
I shall go have a search as my boss hits this a lot, although I tend to see it as cosmic revenge for running Windows when he knows better.
Todays difference betweeen gecko and IE, revealed that IE doesn't vertically align table cells correctly in HTML 4.01 (and presumably other HTML versions). Is this "embrace and disfigure"...
But this is the myth that the public sector can't do anything right, and it is equally stupid.
Verizon are terrible at customer service, and I assume a lot of their broadband business is basically down to limited choice, or the fact that their competitors are worse.
Most of the world telecoms exists due to governments building it, and privatising it when the market was established.
Certainly here (like I believe the US) privatisation failed to break up the effective monpoly postition of the established supplier (who wants N physical phone networks?), who has gone on to profiteer at the publics expense.
So the real question is do these things fit into the category of "natural monopolies", and I think wireless services may do, although probably slightly less clear than for other utility services.
"Is that really what you want to rest your business plan on? A bunch of bickering anti-social primadonnas?"
And worse still most of them have a day job at proprietary software houses, where of course their bickering and internal politics never affects the quality of the end product. And even if it did you'd never get to hear about it till support is dropped. Is that what you want to base your business plan on?
No you have to use/dev/urandom to remove random bits of files under/etc to simulate the "disk filled up whilst applying this service pack despite it claiming to have enough space before hand" feel you get with a registry.
Bitter me --- it was only another two days of my life in servitude to the evil empire.
"Debian is a platform, its simply not a platform targetted at businesses."
I always thought Debian was targeted at system administrators. The packaging in Debian is precisely about sharing administration experience in the same way free software shares the ability to do something with software.
As such Debian is very widely used in businesses, I suspect quite a lot of the top business people don't even know it, they just find the techie guys are asking for less money for software and still turning out better levels of performance.
"From this point of view, an OS that wants to be so universal will never be the best on any single platform. Therefore, Linux will never be better than windows, if we are talking about Inteles alone."
Isn't that the "lets kill a port" myth of Debian release management in another guise, and with less basis.
The platform is pretty well hidden these days, low level programmers occaisonally scratch head when they forget there is more than one number representation.
What will determine success on a platform is down to how well THAT platform is supported, and I don't think being supported well on other platforms is a draw back.
Indeed, in the 90's there were a lot of architectures on which proprietary Unixes (mostly SVR4) were the best and we didn't see Unix vendors going round saying "we must be second best of QUMA because we are Unix and run on lots of platforms.".
Would you want to support Lotus Notes, with their own way of doing everything dating from before most major OSes had things like PPP, and now expected to support everything integrated with the client.
Sell it to CA they know how to let a product die gracefully.
Slashdot Mirror
> 1. The license only restricts your ability to take java, change the name and call it your product, then start charging for it, without paying license and royalty fees to Sun.
Funny I could swear I cut and pasted this from the SUN site.
"Modified source code cannot be distributed without the express written permission of Sun Microsystems, Inc."
So basically you can read the code to find out how it works, but you can't distribute bug fixes, or enhancements, you can't port useful bits of the code to other Java implementations or other software (indeed writing your own version after you've read the source code might be legally risky I suspect). It makes no mention of whether you charge for it.
Indeed the licence attempts to risk redistribution of modified binaries "internally", so modifying the code for your companies own private use may be a licence infringement.
"Here is the source code, look and admire, but don't touch it."
Compare and contrast the licence with the other 16,000 odd packages in Debian.
The article suggests;
;)
Dell's personal worth 1.6 x 10^10
Investment in Redhat 10^8
What proportion did you say again
Don't get me wrong I think your comment is right to the point. But this would be like me getting a mortage of 100,000USD, and then investing 600USD in Redhat.
Because of the numbers involved it sounds big finance, but without knowing the portfolio you can't know if it is hedging, or represents a major personal commitment to Redhat.
Michael Dell is a smart business man, I'm betting he sees it as a long term bet. If you buy the "GNU/Linux will be the next big thing" (and it already is in the server market), and look for a company to invest in to gain from this, Redhat is the obvious choice (IBM is just too big, they could succeed in GNU/Linux big time and only affect the business value by a few points, many of the little exciting companies are just too unpredictable when you have 100 million to invest).
Virus scanning is so obviously not the fix to this problem I'm not sure where to start.
I don't run virus checking on this PC because there are no major viruses in the wild for the operating system, because people deal with the issues properly in a timely fashion when they arise.
Sure it is vulnerable to a new types of attack in a new vulnberability, but virus checking my email is very unlikely to stop that.
Virus checking only protects people who don't deal with issues in a timely fashion, or whose vendor doesn't.
Virus checking for email is basically a short term fudge till people get decent email clients, or update the existing one.
ISPs could refuse to send email for people using old versions of Outlook/Outlook Express. This would be just as effective in stopping email bourne viruses I suspect, and far easier to implement.
Not sure what you are trying to say.
The spam problem viruses tend to set up SMTP proxies that allow anonymous amplification of the injected message. Not a true SMTP server, but not really a conventional client.
"That is properly the smartes idea in this thread."
Till the spammer send out an Official looking Microsoft CD that owns the recipients PC.
Microsoft will send patch CDs out to anyone that asks for them AFAIK, which is the right way to do it. Most good computer magazines also often ship copies of Microsoft patches on cover CDs/DVDs.
At the end of the day the end user still has to take a positive action till SP2 appeared, even if it is "insert CD", which alas it often isn't.
Each of these CD ought to trigger a patch level test and provide a simple "You are out of date?" "Get Upto Date" "I really know what I'm doing" dialog. Because the problem users are mostly the ones who would rarely click on a "I know what I'm doing" button.
Yes but here as pointed out this is not a big problem in terms of numbers.
;)
You see perhaps one or two boxes attempting a brute force ssh password attack a day on a well hosted server.
Sure it creates a lot of log file mess, but that is because by default it logs every attempt, and they often try 200 root passwords in one go.
Dshield data suggests they see 600 sources looking for port 22 a day, if we assume a fair number of these are just port scans, it doesn't leave much activity, and I've seen it suggested this is only semiautomated.
Definitely the worst current problem for GNU/Linux. But fairly stable, so presumably this is mostly people who don't give a monkeys that their box has been owned.
One day someone will get pissed off and try to brute force the passwords on these attackers
I think some GNU/Linux users are complacent to about security, but I doubt we'll ever see anything on the scale of the Sober viruses till there is one big sucessful Desktop distribution.
I just use Postfix MIME type header inspection to reject Windows executable content types.
.tar.gz extension anyway ;)
It is literally a three line change to Postfix, one to say use this file of REGEXP, and two regular expressions. Linked from the postfix.org site.
Do that in advanced of the antivirus, and you'll reduce the load to just unpacking those damn "zip" file attachments.
For my own email I reject ZIP file attachments as well. Realised I was getting many thousands of viruses to every genuine ZIP file. And most stuff that matters was a
"It's interesting because it means that there are still enough unpatched machines out there for a worm to gain serious traction without uncovering new technical vulnerabilities."
You'd have to be living on another planet for this to be news.
Depressingly local computer shops were selling Windows with only SP1 installed a long time after SP2 was release and certainly within the past 6 months.
The answer to "why now" is on the Veritest web page.
The paper is hilarious if you actually read it.
Key points...
They looked for tips on optimising Apache on the Redhat website ?!(guys next time try httpd.apache.org).
They found that Apache 2 was 50% faster than Apache 1 (without any tuning, Apache 2 offers a selection of threading models so a fair comparion would have tried each in turn, with or without tuning), so presumably didn't do any further tests with that in case it made MS look bad.
They tuned Windows for the server but effectively plead ignorance of how to tune Redhat for the server.
My guess is even then, on this hardware, sensible tuning of the kind they did to Windows would have made Apache comparable or better.
There were issues with this hardware selection at the time. Driving Gigabit ethernet is pretty demanding stuff, and you need drivers that can handle interrupt load.
However anyone who actually needs an 8 CPU machine to serve gigabit websites will probably do their own benchmarking and tuning, in the unlikely event their application software gives them a free choice of platform.
When my employers online business is big enough to need gigabit hosting, we'll probably still have SQUID on Linux with shed loads of memory accelerating the static content, because hardware is cheaper than rewriting the corporate applications.
As someone using Unix for nearly 20 years, I'd have to ask also "is the change worth it?".
Whilst there appears to be some gain in consistency in going the launchd way, there is no particular gain in functionality or flexibility over the a combination of initd, and xinetd, and the SVR4(ish) init scripts. Indeed scripts can do everything, and I'd be surprised if some of the things Apple calls aren't scripts precisely to do "extra stuff".
Others have proposed different daemon handling schemes before, such as DJB. Each scheme has its merits, and it is always cleaner to start again.
But it is easy to underestimate the amount of work in moving a big GNU/Linux distro to a different start-up system, not just scripts but documentation as well, and the loss of "know how" as all those thousands of usenet posts no longer answer the question of why o and so application doesn't start on boot.
That said I've never been a fan of the bastardised Linux way of doing start-up scripts. I remember when run levels were really levels.
Inefficiencies of lauching a shell as the slow bit?
Well it is possibly non-trivial I see about 1 sec for launching 100 shells sequentially on my desktop, but I'd always been led to believe the predominant reason for slow deskstop start-up on GNU/Linux was the sheer number of files required to get a working desktop together (~500), which could only be reduced by eliminating seek time (boot from flash), or a radical restructuring of XFree86.
Certainly the last time I saw MACOS X boot it didn't look especially fast.
I vow to always reply to these threads when I see them with my latest stats, just to hack people off ;)
Yesterday 81.1% of hits on our site were IE.
Firefox in second 11.1%.
Safari, "Other Gecko", and the ubiquitous "Unknown" were the only other browser grouping to score more than 1%.
Konqueror barely exceeded wget 458 versus 312, both show as 0.0% when rounded to one decimal place by our stats program.
The OS stats continue to suggest a monopoly position for Microsoft, although I need to dig into that "3.1% unknown" which is the largest single alternative, to see how "Unknown" it really is, and how much I need to update the patterns.
Saying KDE is not cheating, we are talking about Window Manager levels, it would be the appropriate tool to compare to an Apple Desktop.
;)
Or is AppleScript going to magically allow control of all apps written in any combination of languages and tools that can be compiled on Macos X?
But then we are buried beneath a thread where someone seemed to suggest Linux in response to "scripting". Scripting predates me, and I predate Linux somewhat. I've only briefly used Multics in anger, but it felt terribly familiar.
Be assured whatever Apple release it will make "Microsoft Test" look rather feeble
kdcop
I think there are two aspects here.
One is purely performance/hardware, it has long been possible to do a free text search of all the contents of your computer (security issues aside - this is very hard to do securely in a multiuser environment if anyone cares about security/privacy), just no one could be bothered to sacrifice the amount of disk space, and CPU to do this historically. Thus this is purely a technology whose time has come, previously Microsoft and others were mostly indexing metadata (optionally, like you'd bother to switch it off these days, but 5 years ago you may well have chosen to), or select portions of the system.
When Google announced they were doing this I downloaded a free software product that already did this for GNU/Linux for comparison, and yes it worked, it also took about a day to index my system as root, doubled my disk space usage, and needed to refresh its indexes (which it chose to do overnight), and made everyones content visible (if only indirectly) to anyone who could query the database.
As such indexing is a natural progression of computing, but as any database person will tell you indexing has a big resource cost. These days no one cares if their 150GB hard disk is 20% used instead of 10%, and if writing a file takes 10000 operations instead of 100.
The other is integration, Apple as always have a truely innovative integration between search and desktop. I don't always buy their "usability" features (I like menus in the Window they relate to, saves a LOT of mouse mileage). But I think they probably have the edge here, and will retain it if only because of the relatively small application base installed on most Apple desktops.
Perhaps you are overestimating how much impact one department or group at a university has on the others.
I wonder how many Universities had similar compromises and nobody ever noticed?
"IMO that's bad programming"
Problem is it isn't clear whose bad programming. This slow switch only happens on Windows, I suspect as OS/Library bug, but it is possible it is something in the Windows version.
I shall go have a search as my boss hits this a lot, although I tend to see it as cosmic revenge for running Windows when he knows better.
Todays difference betweeen gecko and IE, revealed that IE doesn't vertically align table cells correctly in HTML 4.01 (and presumably other HTML versions). Is this "embrace and disfigure"...
"Of course there are"
But this is the myth that the public sector can't do anything right, and it is equally stupid.
Verizon are terrible at customer service, and I assume a lot of their broadband business is basically down to limited choice, or the fact that their competitors are worse.
Most of the world telecoms exists due to governments building it, and privatising it when the market was established.
Certainly here (like I believe the US) privatisation failed to break up the effective monpoly postition of the established supplier (who wants N physical phone networks?), who has gone on to profiteer at the publics expense.
So the real question is do these things fit into the category of "natural monopolies", and I think wireless services may do, although probably slightly less clear than for other utility services.
"Is that really what you want to rest your business plan on? A bunch of bickering anti-social primadonnas?"
And worse still most of them have a day job at proprietary software houses, where of course their bickering and internal politics never affects the quality of the end product. And even if it did you'd never get to hear about it till support is dropped. Is that what you want to base your business plan on?
Humans - so bloody unreliable....
No you have to use /dev/urandom to remove random bits of files under /etc to simulate the "disk filled up whilst applying this service pack despite it claiming to have enough space before hand" feel you get with a registry.
Bitter me --- it was only another two days of my life in servitude to the evil empire.
"Debian is a platform, its simply not a platform targetted at businesses."
I always thought Debian was targeted at system administrators. The packaging in Debian is precisely about sharing administration experience in the same way free software shares the ability to do something with software.
As such Debian is very widely used in businesses, I suspect quite a lot of the top business people don't even know it, they just find the techie guys are asking for less money for software and still turning out better levels of performance.
"From this point of view, an OS that wants to be so universal will never be the best on any single platform. Therefore, Linux will never be better than windows, if we are talking about Inteles alone."
Isn't that the "lets kill a port" myth of Debian release management in another guise, and with less basis.
The platform is pretty well hidden these days, low level programmers occaisonally scratch head when they forget there is more than one number representation.
What will determine success on a platform is down to how well THAT platform is supported, and I don't think being supported well on other platforms is a draw back.
Indeed, in the 90's there were a lot of architectures on which proprietary Unixes (mostly SVR4) were the best and we didn't see Unix vendors going round saying "we must be second best of QUMA because we are Unix and run on lots of platforms.".
Would you want to support Lotus Notes, with their own way of doing everything dating from before most major OSes had things like PPP, and now expected to support everything integrated with the client.
Sell it to CA they know how to let a product die gracefully.
These days nearly every big general purpose OS implements some or all of the Unix or Posix APIs as it makes it possible to port software.
Not sure that makes them Unix clones, but these days it gets them certified if they want it.
Now back to our regularly scheduled offtopic posts.