I install the OS only once, and installing Solaris is really not that hard.
Drivers in Linux break every now and then - and if your driver is not in the kernel tree you have to compile it yourself. Even if it is, you are the tester. Sun does run tests on Solaris drivers.
NFS is not really part of Linux, unless you refer to some 20 year old version.
Getting updates for Sun packages has never been a problem for me. Yes, an automated system which fetches the updates would be very nice, and so would be faster updater (it is horribly slow). But at least Fedora Core 6 has had working autoupdater for, let's see... zero days (it fails to inform about updates far too often).
Non-Sun packages from Blastwave are as easy as in Linux, although sometimes slightly outdated.
The installation problems are completely immaterial compared to update stability. For example all device drivers keep working after kernel update because they are both tested and binary compatible. This alone makes Solaris far superior compared to Linux (especially in servers). ZFS is a nice addition, and so is DTrace, NFS, etc.
Adding additional software is now about as easy as in Linux, but keeping system up-to-date is not (AFAIK). Hopefully Solaris will improve on both accounts.
Back in -93 UnixWare was the best Unix, on x86 platform.
I still have the CD, maybe I should try it:-) I even have the development books (they were quite expensive, btw).
I was heavily disappointed when Novell abandoned Unix business, there was nothing good left. Linux back then was not really usable, IMHO, although that is what I picked (with Solaris).
Just that Kasparov could only be beaten if they kept changing the goalposts on him. The way Kasparov played he deserved to lose. He would have lost even if the Deep Blue would not have been modified. It was rather not how "good" DB was than how "bad" Kasparov was (on that match).
I was not trying to disagree completely with the GP as clearly JIT'd languages are biggest winners. Just noted that "even" C can be improved dynamically, without compiler help. I would not call it a scheduling mechanism, rather a code morphing mechanism.
Besides the proposed extensions (the PDF) did not so much help in thread scheduling but rather on cache coherency and finding hot spots. Garbage collection can, if it would use the cache information, improve performance perhaps a lot just by moving the objects around. HotSpot VM can use the, eh, hot spot information without counting in SW.
Of course HotSpot-like JIT'd languages are "easiest" target and most likely gives the biggest performance improvement. After all, HotSpot does partially (in SW) what the proposal does (in HW).
If you just CANNOT apply a patch Could it be so that the problem is exactly here? I mean WHY they could not apply. It was claimed that some of the hardware no longer worked. The only reason I can imagine is that some driver got broken and/or was not supported by later kernel.
Hmm... perhaps, just perhaps, this could have been avoidable by a stable binary interface in the kernel... no, I would say it would be more than likely.
It is there enough that I believe if I installed it myself on a poweruser friend's computer, helped sort out some driver issues, he'd be able to take it from there with the occasional internet search. Unfortunately I do not think so. The biggie is need to recompile gspca (webcam) and nVidia and... drivers after every kernel update - which are quite frequent.
I would not put Linux to a friends system whom I am not certain is capable of compiling and installing those drivers - and debug problems which will happen. I think it is above "power user".
Linux really needs a stable KBI (kernel binary interface), no matter what Linus, etc. claim.
A junior programmer (fresh out of school) gets around 75K USD (before taxes) Not even close.
I would estimate an university graduate gets (depending where s/he lives) 2000-3000 euros per month. Seasoned programmer gets between 3000 and 5000. 3000*12.5*1.3 = $50k (12.5 because of summer holiday "extra").
The minimum wage is neither near $32k, it is just above $15k (in practice slightly bigger. This is in Finland, varies from country to country). That is for a full time job - the waitresses and cleaners do not always get one.
Open Document Format lacks some features expected from modern office applications. As a result, many applications use their own extensions. Name one. Don't be fooled to spreadsheet formulas: they all use extremely compatible stuff, see OASIS spec for that.
Office Open XML has a complete feature set. Really? http://atrc.utoronto.ca/index.php?option=com_conte nt§ionid=14&task=view&hidemainmenu=1&id=371. I would not call that "complete".
Office Open XML exactly specifies all of that. You have forgotten the "layoutLikeWord95", etc.
While Open Document Format is an open standard, one usually uses the OpenOffice.org flavour. The flavour is called ODF1.1. What other flavours there are in use?
Open Document Format allows easy upgrading from StarOffice/OpenOffice.org documents. Yes and no. It is equally easy for other applications too.
Office Open XML allows easy upgrading from Microsoft Office documents. But makes it practically impossible for other applications.
Combining the specifications is a very silly proposal, you apparently lack the knowledge of the specifications.
Third parties write crap, exploitable code and it's MS's fault? I think the OP meant that requiring signed drivers is inconvenience to the driver writers and users. A lot like Linux requiring recompile of non-distribution-included drivers in every kernel patch.
Do I need to say why Microsoft likes signed drivers? Do I need to say why Linus likes to break out-of-kernel-tree drivers? Both reasons are equally idiotic, btw.
I have not determined such a thing. See, I find some stuff more reprehensible than other, and somewhere there is a line after which I boycott a company: if I can choose between brand X and brand Y the companies making them may affect my purchase, not only price.
Of course your method "nothing is 100% clean" is a lot easier.
Sure the remount works. But it would be nice if the automounter when mounting would mount it correctly. It seems the automounter (or whatever it actually is) does not read fstab:-(
I checked it with "mount" command, it gives mounts and options when run without arguments (seems to give same information as/etc/mtab gives).
I think your argument ("absence of all Western influence wouldn't bring peace") is... how should I say it... extremely idiotic. Sorry, not intented as an insult.
I think it is our (westerners) responsibility to try to improve situation by bringing "more" peace, not less.
My point tried to be that at least in Finland (and I suspect in a lot of other jurisdictions too) the liability is not same for all: an individual (free or OSS) is less liable than a company (paid software).
IANAL too. The merchantability and fitness for a particular use cannot be disclaimed, not at least for other products. I severely doubt they can be disclaimed for software.
But for used (second hand or self made) items law is different. Unless you sold the product knowing the faults and the purchaser did not know about them (i.e you did not tell him/her) you are not liable.
But for the companies... the product must be fit for the purpose, or you are entitled for at least full refund. What would happen to indirect problems (e.g. SW loses your database) AFAIK nobody knows. OTOH, if e.g. a circular saw cuts your hand off due to a defect in the saw (whether mentioned in the user's guide or not), the manufacturer (or seller or importer) is liable for your medical and disability pension costs. There is no way seller can disclaim this (notice second hand sales - I can sell defective circular saw as long as I state the problems to the prospective buyer).
In EU, and especially in Finland, companies do have a lot of more responsibility. I think it is a good thing.
Your question about "materially negligent"... I am not willing to make any guess where the "line" (for criminal behavior) is. In practice I am very confident no judge would ever give any penalty on free software writer, unless the writer "left" the security problems on purpose (to be used as a trojan/virus/whatever).
Slashdot Mirror
I install the OS only once, and installing Solaris is really not that hard.
... zero days (it fails to inform about updates far too often).
Drivers in Linux break every now and then - and if your driver is not in the kernel tree you have to compile it yourself. Even if it is, you are the tester. Sun does run tests on Solaris drivers.
NFS is not really part of Linux, unless you refer to some 20 year old version.
Getting updates for Sun packages has never been a problem for me. Yes, an automated system which fetches the updates would be very nice, and so would be faster updater (it is horribly slow). But at least Fedora Core 6 has had working autoupdater for, let's see
Non-Sun packages from Blastwave are as easy as in Linux, although sometimes slightly outdated.
The installation problems are completely immaterial compared to update stability. For example all device drivers keep working after kernel update because they are both tested and binary compatible. This alone makes Solaris far superior compared to Linux (especially in servers). ZFS is a nice addition, and so is DTrace, NFS, etc.
Adding additional software is now about as easy as in Linux, but keeping system up-to-date is not (AFAIK). Hopefully Solaris will improve on both accounts.
Route 1: http://opensolaris.org/os/community/on/devref_toc/ devref_1/
Jump to chapter 1.3.3. Follow instructions.
Route 2: http://www.gnusolaris.org/gswiki
Follow instructions.
HTH. HAND.
Back in -93 UnixWare was the best Unix, on x86 platform.
:-)
I still have the CD, maybe I should try it
I even have the development books (they were quite expensive, btw).
I was heavily disappointed when Novell abandoned Unix business, there was nothing good left. Linux back then was not really usable, IMHO, although that is what I picked (with Solaris).
You are right, the reference I gave is old.
I was not trying to disagree completely with the GP as clearly JIT'd languages are biggest winners. Just noted that "even" C can be improved dynamically, without compiler help. I would not call it a scheduling mechanism, rather a code morphing mechanism.
Besides the proposed extensions (the PDF) did not so much help in thread scheduling but rather on cache coherency and finding hot spots. Garbage collection can, if it would use the cache information, improve performance perhaps a lot just by moving the objects around. HotSpot VM can use the, eh, hot spot information without counting in SW.
There already are systems which do exactly that (optimise dynamically C programs), see http://arstechnica.com/reviews/1q00/dynamo/dynamo- 1.html.
Of course HotSpot-like JIT'd languages are "easiest" target and most likely gives the biggest performance improvement. After all, HotSpot does partially (in SW) what the proposal does (in HW).
Hmm
I would not put Linux to a friends system whom I am not certain is capable of compiling and installing those drivers - and debug problems which will happen. I think it is above "power user".
Linux really needs a stable KBI (kernel binary interface), no matter what Linus, etc. claim.
There is WGA, there are numerous viruses (etc), there is this DRM, ...
How badly people are addicted to games? Are they willing to give up their first born - where's the line?
Written by a totally flabbergasted non-gaming individual.
I would estimate an university graduate gets (depending where s/he lives) 2000-3000 euros per month. Seasoned programmer gets between 3000 and 5000. 3000*12.5*1.3 = $50k (12.5 because of summer holiday "extra").
The minimum wage is neither near $32k, it is just above $15k (in practice slightly bigger. This is in Finland, varies from country to country). That is for a full time job - the waitresses and cleaners do not always get one.
I live in Finland, I am a (seasoned) programmer.
I do not think mortgage payments are deductible, but the interest is. At least his is how it works in Finland, so I could be wrong.
Combining the specifications is a very silly proposal, you apparently lack the knowledge of the specifications.
I do. Not only mind, but object, severely.
e nt§ionid=14&task=view&hidemainmenu=1&id=371.
s
l e.php?story=20070801182558375. Paid shills.
One of the reasons I object is: http://atrc.utoronto.ca/index.php?option=com_cont
There are a lot of other technical problems why it should not be a standard listed in various places on the web, see e.g. http://www.grokdoc.net/index.php/EOOXML_objection
BTW, I am disappointed on the disabled community for not "standing up" - as they did with ODF 1.0 (and therefore ODF 1.1 was created) http://www.consortiuminfo.org/standardsblog/artic
I'll respect Microsoft the day they respect me.
Just follow, for example, ODF v.s. EOOXML.
Do I need to say why Microsoft likes signed drivers? Do I need to say why Linus likes to break out-of-kernel-tree drivers? Both reasons are equally idiotic, btw.
I have not determined such a thing. See, I find some stuff more reprehensible than other, and somewhere there is a line after which I boycott a company: if I can choose between brand X and brand Y the companies making them may affect my purchase, not only price.
Of course your method "nothing is 100% clean" is a lot easier.
Sure the remount works. But it would be nice if the automounter when mounting would mount it correctly. It seems the automounter (or whatever it actually is) does not read fstab :-(
/etc/mtab gives).
I checked it with "mount" command, it gives mounts and options when run without arguments (seems to give same information as
How do I disable atime from USB disks? The fstab method does not work, not at least with Fedora Core 6.
I think your argument ("absence of all Western influence wouldn't bring peace") is ... how should I say it ... extremely idiotic. Sorry, not intented as an insult.
I think it is our (westerners) responsibility to try to improve situation by bringing "more" peace, not less.
Virtual machine is perhaps too extreme, but how about http://docs.freebsd.org/44doc/papers/jail/jail.htm l?
There is also java.policy in jre/lib/security directory, but I doubt you can disable the system modal from there.
I fully agree with your answer.
My point tried to be that at least in Finland (and I suspect in a lot of other jurisdictions too) the liability is not same for all: an individual (free or OSS) is less liable than a company (paid software).
IANAL too. The merchantability and fitness for a particular use cannot be disclaimed, not at least for other products. I severely doubt they can be disclaimed for software.
... the product must be fit for the purpose, or you are entitled for at least full refund. What would happen to indirect problems (e.g. SW loses your database) AFAIK nobody knows. OTOH, if e.g. a circular saw cuts your hand off due to a defect in the saw (whether mentioned in the user's guide or not), the manufacturer (or seller or importer) is liable for your medical and disability pension costs. There is no way seller can disclaim this (notice second hand sales - I can sell defective circular saw as long as I state the problems to the prospective buyer).
... I am not willing to make any guess where the "line" (for criminal behavior) is. In practice I am very confident no judge would ever give any penalty on free software writer, unless the writer "left" the security problems on purpose (to be used as a trojan/virus/whatever).
But for used (second hand or self made) items law is different. Unless you sold the product knowing the faults and the purchaser did not know about them (i.e you did not tell him/her) you are not liable.
But for the companies
In EU, and especially in Finland, companies do have a lot of more responsibility. I think it is a good thing.
Your question about "materially negligent"