Where went the albums? Back in seventies there were *albums*, rather than songs.
You would listen an album of related music (in a kinda "story") rather unrelated list hits from various artists.
I think this is one thing which is wrong with the modern music: they are filling CDs with songs that might include one hit No continuity, no nothing, just unrelated perhaps-a-hit stuff ending up as likely-a-shit.
First, it does not need to be carved in stone. Change it in every major version (e.g. 2.4 -> 2.6).
Second, if refactoring needs to be "normal" part of Linux development it just tells how horrible shape Linux is in.
Third, Solaris has had quite a stable ABI since 2.5.1 (and Windows since w2k) and it has not (much) hindered development - I would not say that is "much slower". In some parts it is faster than in Linux.
DRM. You forgot DRM. I would add example of how e.g. HDCP makes consumers pay more (e.g. I have DVI capable TV - but no HDCP). And, actually, so does every other DRM scheme known to man.
You forgot safety *for the taxi driver*. At least in Finland this is seen as a safety feature: if someone tries to kidnap and/or rob a cab driver this, among with other features, will help.
There were a couple of dead taxi drivers because some junkies robbed them... after GPS and cameras none.
2) Apparently financial calculations are forbidden by OOXML.
6) The Linux way of doing it is wrong, too. It makes *correct* translation of some idioms far too difficult. I do not claim there is a perfect solution for this - there isn't.
7) Think about security issues. Using this (and perhaps a few other flaws) I could probably make a spreadsheet which changes values e.g. depending on when viewed. Without using macros or functions.
13) Isn't it a marvellous standard, this Ecma 376? No implementation ever will use it. Why standardize it, then?
Your point that the problems are in Excel, not in Ecma 376 is 100% beside the point as Excel defines the OOXML, not the Ecma standard. See point 13.
http://opensource.org/licenses/alphabetical has 59 licenses, a couple of which are "not recommended" anymore (at least "Historical Permission Notice and Disclaimer", "SISSL" and "X.Net"). Some of them are pretty much identical to each other (near public domain), some are "version two".
So it would make around fifty different "alive and well" licenses.
So when do you think Microsoft, sorry Ecma, will correct the accessibility problems? You see, ODF1.0 is the ISO standard, but everybody agrees ODF1.1 is the specification to use. It corrects a few accessibility problems, unlike Microsoft.
Or perhaps you can show me a link to a open process by Microsoft which tackles that (and the zillion other problems shown e.g. in grokdoc).
Because you do know the link to the open process tackling the formula "problem".
The article completely misses an important fact: most of the files on the disk never change.
Suppose I have a 100GB disk and 99GB is full of non-changing data. Then the changing data can only use 1G reaching the 100k-2M write cycles couple of orders magnitude faster than the article claims.
Yes, my calculation is certainly "worst case" (you should never fill disk over 90% full), but it shows that the article is wrong (it claims worst case).
However, the conclusion is correct: you need not worry about the wear.
Hmm... how about a system which automatically compiles and optimises the program for the processor where the program is run? The program writer could send the program in some intermediate language, let's call it "byte code". The VM^H^Hsystem could even figure out what parts of the program to optimize based on use kinda "searching" for "hot spots"?
Of course it is! It would take just a few minutes for me to generate practically infinite number of either... so I can decide with 100% accuracy which one is more relevant! Pay me to decide the matter for good!:-)
Slashdot Mirror
Where went the albums? Back in seventies there were *albums*, rather than songs.
You would listen an album of related music (in a kinda "story") rather unrelated list hits from various artists.
I think this is one thing which is wrong with the modern music: they are filling CDs with songs that might include one hit No continuity, no nothing, just unrelated perhaps-a-hit stuff ending up as likely-a-shit.
You are too kind. You both(?) forget that the "Fast Track is not a standards development process" http://www.robweir.com/blog/2007/09/how-to-hack-iso.html.
Besides, there are much worse problems than the "xxxLikeWord95", see e.g. http://en.wikipedia.org/wiki/OOXML#Technical_criticisms.
Furthermore I am quite certain Microsoft Office 2007 will never support OOXML as specified.
If we all pull together we might, just might, solve Global Warming. But the other one ... it's like understanding women. No chance.
DRM. I am certain the Linux version will not play DRM'd content.
First, it does not need to be carved in stone. Change it in every major version (e.g. 2.4 -> 2.6).
Second, if refactoring needs to be "normal" part of Linux development it just tells how horrible shape Linux is in.
Third, Solaris has had quite a stable ABI since 2.5.1 (and Windows since w2k) and it has not (much) hindered development - I would not say that is "much slower". In some parts it is faster than in Linux.
There went the credibility of the ISO in one go.
Forward compatible device drivers, i.e. (more or less) stable KBI.
I REALLY hate to compile gspca and nVidia after every kernel update (which there are, IMHO, too many in every OS known to man).
Whether this is good enough for you or someone else is a personal matter. It certainly is good enough for me.
Why cannot it get the driver via network (Vista can from USB, I think)?
WTF? They did exactly same in Germany, Australia, Hungary, Switzerland, Norway, ... How many there must be before it is "company policy"?
Molehill? Please ...
Maybe, just maybe, they are only avoiding, consciously or unconsciously, those with a bad name. I know I would.
Let's make it sane: use some distributed version control systems which do merging so much easier.
DRM. You forgot DRM. I would add example of how e.g. HDCP makes consumers pay more (e.g. I have DVI capable TV - but no HDCP). And, actually, so does every other DRM scheme known to man.
You forgot safety *for the taxi driver*. At least in Finland this is seen as a safety feature: if someone tries to kidnap and/or rob a cab driver this, among with other features, will help.
... after GPS and cameras none.
There were a couple of dead taxi drivers because some junkies robbed them
2) Apparently financial calculations are forbidden by OOXML.
6) The Linux way of doing it is wrong, too. It makes *correct* translation of some idioms far too difficult. I do not claim there is a perfect solution for this - there isn't.
7) Think about security issues. Using this (and perhaps a few other flaws) I could probably make a spreadsheet which changes values e.g. depending on when viewed. Without using macros or functions.
13) Isn't it a marvellous standard, this Ecma 376? No implementation ever will use it. Why standardize it, then?
Your point that the problems are in Excel, not in Ecma 376 is 100% beside the point as Excel defines the OOXML, not the Ecma standard. See point 13.
Why didn't you count?
http://opensource.org/licenses/alphabetical has 59 licenses, a couple of which are "not recommended" anymore (at least "Historical Permission Notice and Disclaimer", "SISSL" and "X.Net"). Some of them are pretty much identical to each other (near public domain), some are "version two".
So it would make around fifty different "alive and well" licenses.
I do not think the "UseWord95Spacing" is as big a problem as it has been made to look. Much worse are other aspects, like problems with accessibility http://atrc.utoronto.ca/index.php?option=com_conte nt§ionid=14&task=view&hidemainmenu=1&id=371 and problems with robustness http://www.arstdesign.com/articles/DefectiveByDesi gn.htm.
The blog to follow is Rob Weir's.
So when do you think Microsoft, sorry Ecma, will correct the accessibility problems? You see, ODF1.0 is the ISO standard, but everybody agrees ODF1.1 is the specification to use. It corrects a few accessibility problems, unlike Microsoft.
Or perhaps you can show me a link to a open process by Microsoft which tackles that (and the zillion other problems shown e.g. in grokdoc).
Because you do know the link to the open process tackling the formula "problem".
Aren't the open processes marvellous?
The grokdoc misses at least the biggie at http://atrc.utoronto.ca/index.php?option=com_conte nt§ionid=14&task=view&hidemainmenu=1&id=371.
The article completely misses an important fact: most of the files on the disk never change.
Suppose I have a 100GB disk and 99GB is full of non-changing data. Then the changing data can only use 1G reaching the 100k-2M write cycles couple of orders magnitude faster than the article claims.
Yes, my calculation is certainly "worst case" (you should never fill disk over 90% full), but it shows that the article is wrong (it claims worst case).
However, the conclusion is correct: you need not worry about the wear.
Yes, (2^15)-1 is for the positive (0x7FFF), but you do get "full" -2^15 (0x8000) for negative. It is two's complement, not one's.
:-)
This must be one of the most nitpicking posts ever
Hmm ... how about a system which automatically compiles and optimises the program for the processor where the program is run? The program writer could send the program in some intermediate language, let's call it "byte code". The VM^H^Hsystem could even figure out what parts of the program to optimize based on use kinda "searching" for "hot spots"?
Wouldn't something like that sound marvellous?
I think the requirements are a bit low, if you compare them to http://www.tomshardware.com/2007/07/16/cpu_charts_ 2007/page21.html.
But then, the content is somewhat different.
But not HD. Most likely the bit rate is limited to 10Mbit/s and resolution (on disk) to 720x576.
For HD it should be able to handle at least 20Mbit/s and 1280x720 resolution.
Of course it is! It would take just a few minutes for me to generate practically infinite number of either ... so I can decide with 100% accuracy which one is more relevant! Pay me to decide the matter for good! :-)
Every single Linux fanboy always forgets testing.