We use Mozilla as the Browser+Mail app at work. So I know about some "issues" (bugs) that have shown, and searched Bugzilla for them. Often they are in Bugzilla, and there is some thread of discussion and the notion that they should be fixed. E.g. the very slow operation of IMAP, even when the server is very fast.
I also reported two bugs myself. One was a cosmetic bug that attracted some attention, and is now in state NEW. Another one, that is much more important, is still sitting in state UNCONFIRMED after a month, although it has a very simple and clear reproduction scenario.
It does not seem the above bugs will be fixed in 1.4, and it also seems that will be the final release (or will there be a 1.4.1 etc to fix problems that are not showstoppers?)
With so many bugs still open (also mentioned by others), what use it is to test these RC versions and report more bugs? I can report a couple of bugs offhand, but I am reluctant to spend a lot of time to do the dup research and to spell out clear reports and scenarios, when there is no real chance that the bugs will ever be fixed.
>Yeah, but have you ever called the tech support line for a company who does this? I don't know about you, but I can hardly understand what the techie is saying. I mean no offense to the Indian readers of Slashdot, but some of the accents I just can not understand. I think there are a lot of people [in the US] who are the same way.
There are also many people outside of the US, not native English speakers, who have the same problem when calling a support person in or from the US. Some of the accents are just very difficult to understand!
Any method of detecting certain classes of messages and then discarding them is susceptible to falsely deleting messages. When you don't want that, don't use it. Blacklists also reject valid mail. Especially those blacklists that use the "when they don't want to listen to our complaint we will block their entire network - that will do it" approach. But the lists that have multilevel relays are no better.
It is everyone's own decision. I get about 90% spam, and all the filtering makes me see almost nothing. I peek in the Junk folder to see if any likely false positives are there, and empty it.
When you are so desperate for that single message that may save your life, stay away from filtering. When your life is ruined by 100 SPAM messages a day, that may be a higher priority.
And when that happens, the standards people will only be able to blaim THEMSELVES! They have stood at the line for 10 years and watched the demise of the mail system, bickering about the potential problems of solutions that people proposed.
I would not be surprised when Microsoft comes up with a completely incompatible "enhanced spam-free e-mail system" comprised of only Microsoft products and de-facto phasing out the existing system over a period of one or two years. Just like they did with the webbrowser.
Of course whole/. will be shouting, but what have they done to prevent it?
This does not seem to be warranted... With Linux, I gave gone through extfs, ext2fs and reiserfs. I still have to lose a byte due to filesystem failure. Others have had problems with early versions, but I have been using the stuff as released with distributions, and it seems it has always been stable enough. You would expect (?) that a new FS in Windows is also tested well enough before being released. Maybe you'll have a problem when you rip an early beta off the net?
When you are concerned that your learning about Solaris is wasting your time when you really wanted to learn about Linux, you are probably wasting your time taking courses. Any courses. Studying should provide you with insight in the matter, and train you to find information. When you are only trying to learn magic spells, and are lost when it suddenly is "useradd" instead of "adduser", you will find yourself at a dead end every 3 years.
SMS Spam is not really a problem here. However, what has become a big problem very quickly: paid SMS "services". You "subscribe" by sending a message to a (usually 4-digit) number, and from then on you receive "information" by SMS, which costs you a certain amount of money (around 1 euro) per received message. The messages you receive usually don't have a clear source number, and it is not clear how to stop the stream. It is also unclear how many messages you will receive.
It was often very unclear from the advertisements that one would subscribe to a service. E.g. an advertisement says "receive your horoscope by sending 'yes' to 1234" (1 euro/message) but it did not tell you that this would mean you would receive such a message every day or every week, until you found out how to stop it. And there would be no way to find out, as the message had no source number and the telephone provider would assure you that they were unable to tell who sent you these messages, or to stop them from doing that. (although they of course were able to transfer your money to said party)
After this went way out of control this year, there has been pressure from consumers to regulate these services. It now seems there are some guidelines, to which the providers have agreed to comply. Like publication of the unsubscribe info.
In a network, this seems to be largely redundant. Use PXE when you want a diskless boot. May take more than 3 seconds, but is supported on many, many more systems!
Now you are talking about the computer systems. But the posting was about the building being renovated. It was put up in an extreme hurry, and it is public knowledge that they had big problems with the changed technology before the renovation. (like electrical power to the offices being insufficient for modern offices. of course it was designed for pencil-and-typewriter offices)
This could maybe work for google.com, but not in general. The idea of having addresses and names is that an address is semi-permanently bound to a name, and that it can change when the host moves. Just as your street address. What you propose is akin to naming the street where you live "gfody street". That may be convenient to you, but not to the rest of the world. When you would move, two streets would have to be renamed, and all maps would have to be updated to find the new gfody street.
Not true at all! The "deltree" command, standard with Windows 9x, deletes a tree of directories. The/y flag suppresses warnings. So, deltree/y dirname deletes the directory "dirname".
Now, you might think that this "consumer oriented OS" guards you against mistakes. Not so. When you just type "deltree/y", without specifying a directory, it defaults to the current working directory!!!
Open a command prompt and type deltree/y. Gone is your entire system!! The current working directory when opening a command prompt is (by default) the Windows directory.
Where in the above do you find any guarding agains accidental mis-typing? As an actual fact, this happened to me when there was a line like: deltree/y %VAR% in a login script. The %VAR% was supposed to contain a directory name to clean out. For some reason it was sometimes empty, and in that case the entire C:\Windows directory on the victim's system was erased, less some dlls that were open at that time and could not be deleted. (yet another end-user unfriendly property of Windows, the one that makes it require so many reboots)
>I'm not at all sure if Linux could pass, since it has, eg, a rename(2) system call in place of unlink.
Not in place of, but in addition to. The Unix way of renaming files is to make a new link() to them and then unlink() the old name. However, this assumes the filesystem is able to make multiple links to files.
The Linux system is so versatile that it supports many filesystems, not all of them supporting link(). This is because those filesystems originate from other operating systems where the notion of having multiple links to a single file is usually not supported. To be able to rename files on all filesystems, you need a rename() systemcall.
>Not something for extended type, geosyncronous/etc flights like someone else mentioned. I think you'll need something with a tad more precision and "predicability" for more intense missions
Why do you think that is so much different? In an extended mission you at least have the time to correct and perfect the software, which is not possible in a 10 second duration flight...
To me it seems that the difference is mainly in the hardware, which must be suitable for space when you launch into space and stay there for a while.
Or did you think those designs would not include a reset mechanism and a bootloader?
Actually, what you are describing (user can do nothing, the one and only administrator can do everything) is very easy to accomplish in Linux. It is the default situation.
It would be more difficult when you asked about complicated structures of user groups, where users by being member of such a group can get access to a certain limited number of administrative functions. (e.g. one group can install software, another group can create new users, a third group can change rights on files, etc)
That is what I said: you need a 256MB system. People deny it, but after some discussion they admit that they have only little experience with 64MB and that they upgraded their system because it was sluggish.
That is what I meant with "it is UNUSABLE" with 64MB. People need their systems to be productive, not sluggish. And a 233 MHz 64MB system with KDE 3.0 and OpenOffice plus Mozilla will be extremely sluggish, I can tell you. To the point of being unusable.
The "buy more RAM" idea seems reasonable, however systems of that vintage often do not allow more than 64MB of RAM to be installed and be cached in the 2nd level cache. Also, in an organization with lots of old systems that either have to be replaced or to be upgraded yet again, the decision process is often a bit different than at home.
(I do have 1GB of RAM in my home system, as I know that it is inexpensive and I don't care about investing a little money in my system. Of course the situation is different when you have thousands of systems to do)
>Also, when MS Office is installed, it loads a component that starts every time you start your computer. Be sure that is disabled when you run tests.
Why? When that makes Office start faster, then why would one disable it? Just compare systems as they would be used on the desktop, including any clever tweaks that make applications run faster.
Had you tried OpenOffice, you would have found that it was slower than MS Office just like Netscape/Mozilla is slower than IE.
All these effects combined make that a 64MB machine is usable with Windows and too small for Linux. or better: too small for Mozilla and OpenOffice. We have machines with Windows and those two apps, and this is not workable either. However for Linux on the desktop they are the only real choice.
Well, I don't agree with that. At least not when talking about Linux for the desktop.
A 233 MHz system with 64MB RAM can be used with Windows 9x and Office. It will be UNUSABLE with Linux, KDE or Gnome, and OpenOffice. A newer class of system, say 500 MHz and 256 MB RAM is required for that. Similar to what you would need for Windows 2000.
I'm working in a similar organisation, and it also has way more desktop PCs than you would expect. The situation is currently being investigated, as a replacement of all machines is going to cost too much.
I think there are some simple explanations:
- when newer systems came in, older systems often moved to desks of people not needing a PC. Sometimes saying "it is good when you train yourself in PC use". User now only uses it for mail and games.
- there have been times when PCs came up and everyone needed to have one. It became a status symbol, and not having a PC while someone else had was clearly a difference in ranking.
- in such environments, there often are budgets for these things, that will disappear when not used up, and then decreased for the next year. so at the end of a year, systems are bought just to deplete the budget, and then end up in places where they are not really required.
However, after a couple of years nobody remembers the actual history behind a PC on a certain desk, and it gets replaced by a brand new one at quite some cost (totalling over all PC replacements). So, a good look at the requirements is really needed when a program like this is run.
(I assume that the "so many users to switch to Linux" stories always have the underlying story that all PCs will be replaced and the new systems will run Linux instead of Windows. When this is not the case, they will certainly be disappointed as running a modern Linux desktop and applications on a 2-3 year old system will be a disaster!)
That is easy. It just takes mail from the owner's mailbox and sends it to mail addresses found on the machine. So, when somebody discusses specific topics with a small group of mail recepients, chances are quite high that this happens. (I have seen an example of this as well)
Slashdot Mirror
We use Mozilla as the Browser+Mail app at work.
So I know about some "issues" (bugs) that have shown, and searched Bugzilla for them.
Often they are in Bugzilla, and there is some thread of discussion and the notion that they should be fixed. E.g. the very slow operation of IMAP, even when the server is very fast.
I also reported two bugs myself. One was a cosmetic bug that attracted some attention, and is now in state NEW. Another one, that is much more important, is still sitting in state UNCONFIRMED after a month, although it has a very simple and clear reproduction scenario.
It does not seem the above bugs will be fixed in 1.4, and it also seems that will be the final release (or will there be a 1.4.1 etc to fix problems that are not showstoppers?)
With so many bugs still open (also mentioned by others), what use it is to test these RC versions and report more bugs? I can report a couple of bugs offhand, but I am reluctant to spend a lot of time to do the dup research and to spell out clear reports and scenarios, when there is no real chance that the bugs will ever be fixed.
What do others think?
>Yeah, but have you ever called the tech support line for a company who does this? I don't know about you, but I can hardly understand what the techie is saying. I mean no offense to the Indian readers of Slashdot, but some of the accents I just can not understand. I think there are a lot of people [in the US] who are the same way.
There are also many people outside of the US, not native English speakers, who have the same problem when calling a support person in or from the US.
Some of the accents are just very difficult to understand!
Any method of detecting certain classes of messages and then discarding them is susceptible to falsely deleting messages. When you don't want that, don't use it.
Blacklists also reject valid mail. Especially those blacklists that use the "when they don't want to listen to our complaint we will block their entire network - that will do it" approach. But the lists that have multilevel relays are no better.
It is everyone's own decision. I get about 90% spam, and all the filtering makes me see almost nothing. I peek in the Junk folder to see if any likely false positives are there, and empty it.
When you are so desperate for that single message that may save your life, stay away from filtering.
When your life is ruined by 100 SPAM messages a day, that may be a higher priority.
And when that happens, the standards people will only be able to blaim THEMSELVES!
/. will be shouting, but what have they done to prevent it?
They have stood at the line for 10 years and watched the demise of the mail system, bickering about the potential problems of solutions that people proposed.
I would not be surprised when Microsoft comes up with a completely incompatible "enhanced spam-free e-mail system" comprised of only Microsoft products and de-facto phasing out the existing system over a period of one or two years.
Just like they did with the webbrowser.
Of course whole
Funny that after all this you cannot spell the name of the attacking trojan correctly...
But NTFS also has that...
This does not seem to be warranted...
With Linux, I gave gone through extfs, ext2fs and reiserfs. I still have to lose a byte due to filesystem failure.
Others have had problems with early versions, but I have been using the stuff as released with distributions, and it seems it has always been stable enough.
You would expect (?) that a new FS in Windows is also tested well enough before being released. Maybe you'll have a problem when you rip an early beta off the net?
When you are concerned that your learning about Solaris is wasting your time when you really wanted to learn about Linux, you are probably wasting your time taking courses. Any courses.
Studying should provide you with insight in the matter, and train you to find information. When you are only trying to learn magic spells, and are lost when it suddenly is "useradd" instead of "adduser", you will find yourself at a dead end every 3 years.
SMS Spam is not really a problem here.
However, what has become a big problem very quickly: paid SMS "services".
You "subscribe" by sending a message to a (usually 4-digit) number, and from then on you receive "information" by SMS, which costs you a certain amount of money (around 1 euro) per received message.
The messages you receive usually don't have a clear source number, and it is not clear how to stop the stream. It is also unclear how many messages you will receive.
It was often very unclear from the advertisements that one would subscribe to a service. E.g. an advertisement says "receive your horoscope by sending 'yes' to 1234" (1 euro/message) but it did not tell you that this would mean you would receive such a message every day or every week, until you found out how to stop it. And there would be no way to find out, as the message had no source number and the telephone provider would assure you that they were unable to tell who sent you these messages, or to stop them from doing that.
(although they of course were able to transfer your money to said party)
After this went way out of control this year, there has been pressure from consumers to regulate these services. It now seems there are some guidelines, to which the providers have agreed to comply. Like publication of the unsubscribe info.
In a network, this seems to be largely redundant.
Use PXE when you want a diskless boot. May take more than 3 seconds, but is supported on many, many more systems!
Now you are talking about the computer systems.
But the posting was about the building being renovated. It was put up in an extreme hurry, and it is public knowledge that they had big problems with the changed technology before the renovation.
(like electrical power to the offices being insufficient for modern offices. of course it was designed for pencil-and-typewriter offices)
This could maybe work for google.com, but not in general.
The idea of having addresses and names is that an address is semi-permanently bound to a name, and that it can change when the host moves.
Just as your street address. What you propose is akin to naming the street where you live "gfody street". That may be convenient to you, but not to the rest of the world. When you would move, two streets would have to be renamed, and all maps would have to be updated to find the new gfody street.
Exactly the same happens in an IP network.
Not true at all! /y flag suppresses warnings. /y dirname deletes the directory "dirname".
/y", without specifying a directory, it defaults to the current working directory!!!
/y .
/y %VAR%
The "deltree" command, standard with Windows 9x, deletes a tree of directories. The
So, deltree
Now, you might think that this "consumer oriented OS" guards you against mistakes. Not so.
When you just type "deltree
Open a command prompt and type deltree
Gone is your entire system!!
The current working directory when opening a command prompt is (by default) the Windows directory.
Where in the above do you find any guarding agains accidental mis-typing?
As an actual fact, this happened to me when there was a line like:
deltree
in a login script. The %VAR% was supposed to contain a directory name to clean out.
For some reason it was sometimes empty, and in that case the entire C:\Windows directory on the victim's system was erased, less some dlls that were open at that time and could not be deleted.
(yet another end-user unfriendly property of Windows, the one that makes it require so many reboots)
Try "rmdir /s /q \" on Windows NT and derivatives, or "deltree /y \" on Windows 9x.
(of course there is no cross-Windows-platform way to do it)
>I'm not at all sure if Linux could pass, since it has, eg, a rename(2) system call in place of unlink.
Not in place of, but in addition to.
The Unix way of renaming files is to make a new link() to them and then unlink() the old name. However, this assumes the filesystem is able to make multiple links to files.
The Linux system is so versatile that it supports many filesystems, not all of them supporting link(). This is because those filesystems originate from other operating systems where the notion of having multiple links to a single file is usually not supported.
To be able to rename files on all filesystems, you need a rename() systemcall.
I would have written it when you wouldn't have done so already...
It has always amazed me how the Unix world has constantly been damaging their own business and letting others in, by fighting legal wars.
>Not something for extended type, geosyncronous/etc flights like someone else mentioned. I think you'll need something with a tad more precision and "predicability" for more intense missions
Why do you think that is so much different? In an extended mission you at least have the time to correct and perfect the software, which is not possible in a 10 second duration flight...
To me it seems that the difference is mainly in the hardware, which must be suitable for space when you launch into space and stay there for a while.
Or did you think those designs would not include a reset mechanism and a bootloader?
Actually, what you are describing (user can do nothing, the one and only administrator can do everything) is very easy to accomplish in Linux. It is the default situation.
It would be more difficult when you asked about complicated structures of user groups, where users by being member of such a group can get access to a certain limited number of administrative functions.
(e.g. one group can install software, another group can create new users, a third group can change rights on files, etc)
That is what I said: you need a 256MB system.
People deny it, but after some discussion they admit that they have only little experience with 64MB and that they upgraded their system because it was sluggish.
That is what I meant with "it is UNUSABLE" with 64MB. People need their systems to be productive, not sluggish. And a 233 MHz 64MB system with KDE 3.0 and OpenOffice plus Mozilla will be extremely sluggish, I can tell you. To the point of being unusable.
The "buy more RAM" idea seems reasonable, however systems of that vintage often do not allow more than 64MB of RAM to be installed and be cached in the 2nd level cache. Also, in an organization with lots of old systems that either have to be replaced or to be upgraded yet again, the decision process is often a bit different than at home.
(I do have 1GB of RAM in my home system, as I know that it is inexpensive and I don't care about investing a little money in my system. Of course the situation is different when you have thousands of systems to do)
>Also, when MS Office is installed, it loads a component that starts every time you start your computer. Be sure that is disabled when you run tests.
Why? When that makes Office start faster, then why would one disable it?
Just compare systems as they would be used on the desktop, including any clever tweaks that make applications run faster.
Had you tried OpenOffice, you would have found that it was slower than MS Office just like Netscape/Mozilla is slower than IE.
All these effects combined make that a 64MB machine is usable with Windows and too small for Linux.
or better: too small for Mozilla and OpenOffice. We have machines with Windows and those two apps, and this is not workable either. However for Linux on the desktop they are the only real choice.
Well, I don't agree with that. At least not when talking about Linux for the desktop.
A 233 MHz system with 64MB RAM can be used with Windows 9x and Office. It will be UNUSABLE with Linux, KDE or Gnome, and OpenOffice.
A newer class of system, say 500 MHz and 256 MB RAM is required for that. Similar to what you would need for Windows 2000.
I'm working in a similar organisation, and it also has way more desktop PCs than you would expect.
The situation is currently being investigated, as a replacement of all machines is going to cost too much.
I think there are some simple explanations:
- when newer systems came in, older systems often moved to desks of people not needing a PC. Sometimes saying "it is good when you train yourself in PC use". User now only uses it for mail and games.
- there have been times when PCs came up and everyone needed to have one. It became a status symbol, and not having a PC while someone else had was clearly a difference in ranking.
- in such environments, there often are budgets for these things, that will disappear when not used up, and then decreased for the next year. so at the end of a year, systems are bought just to deplete the budget, and then end up in places where they are not really required.
However, after a couple of years nobody remembers the actual history behind a PC on a certain desk, and it gets replaced by a brand new one at quite some cost (totalling over all PC replacements).
So, a good look at the requirements is really needed when a program like this is run.
(I assume that the "so many users to switch to Linux" stories always have the underlying story that all PCs will be replaced and the new systems will run Linux instead of Windows. When this is not the case, they will certainly be disappointed as running a modern Linux desktop and applications on a 2-3 year old system will be a disaster!)
>There were five topic icons for this story:
Better than five topics for this story!
That is easy. It just takes mail from the owner's mailbox and sends it to mail addresses found on the machine. So, when somebody discusses specific topics with a small group of mail recepients, chances are quite high that this happens.
(I have seen an example of this as well)