It is an interesting question but I disagree with you.
The judge has said that to have a patent being enforced, you must be showing goodwill efforts towards bringing a patent to market. That doesn't cut out small startups, all it cuts out (apart from patent trolls) are defensive patent portfolios and small research houses which attempt to develop and then sell ideas. I think they will be able to adapt to this - show they've sold other patents perhaps. Defensive patent portfolios... well, they're mainly to stop small players and patent trolls - I think weakening them is not such a bad idea.
I was very interested in how well supported the hardware is or whether dell uses binary blobs. Here we are slowly shifting away from windows to ubuntu on the desktop and we already use Fedora on servers - all on dell hardware - so it will be very helpful if it is well supported hardware.
Is that a troll? Would help if you back up the assertion with something (more than IMHO). Personally I don't care who gets the deal as long as it results in roughly standard-compliant comptuers - e.g. we see more of Dell pushing ATI for decent drivers. More support for decent BIOSs, better ACPI.. there are heaps of areas where a large commodity player could help simply by offering more sales.
Oh, and to copy your trollish stance, I hope Ubuntu gets the deal:-P
I deal with a different problem (replace web by database), but my solution might be useful to you.
I feed all logs through a log analyser (implemented using crm114.sf.net). Then the log itself is discarded and only summary statistics are kept. That way you can keep track of both longer term trends and detect when something has gone wrong.
It's a bugger to set up, but if you don't change systems that often is probably worth it.
No, nothing like D&D. It is battle reenactment except it also has fantasy sets (hence orcs) and futuristic sets (robots). There isn't any role-playing or characters - the players are sorta army generals trying to coordinate their armies.
It has the most awkward, painful table storage format. ProC just sucks Installation will leave you tearing your hair out. Configuration is a black art.
There are many good things that can be said about Oracle - the use of indicies inside an index so you don't have to create really unnatural composite indicies... tidy SQL syntax... support for R trees of any number of dimensions... Effective subqurery optimisations... Triggers that work (compared to MySQL, not postgres). Ease of use is nowhere near the list.
AFAICT, they're both about as ugly as each other for writing your own UDFs, my point was I can download a few hundred UDFs for MySQL but I don't seem to be able to do the same thing for Postgress.
--
But say I wanted to give postgres another whirl (mostly to see if it is faster than MySQL at what I normally do). I've found SQL Fairy, is that the best approach?
I don't tend to write SQL directly. I write python code that calls the company database wrapper, which calls MySQLdb, which calls DBI, which calls MySQL. Presumably shifting to postgres involves changing the wrapper to call PyPgSQL. Is that any good, or am I in for a world of pain? I use server side cursors extensively and query the tables, but otherwise most code is just passed straight to the interpreter.
PS: I don't suppose you know if R trees are being extended in postgres beyond three dimensions? If so that would be enough justification to make me want to put serious effort in.
I'm in a similar situation except I use MySQL. Lets look at why.
1) Ease of use: MySQL has readline, and supports lazier SQL programming. It has loads of syntactic sugar that lets you write simple things in whatever way makes sense to you. Adding users, changing passwords, backing up... it is all _easy_. Of course, postgres can do everything MySQL can too, the question is which is easier.
2) Extensive developer community. We use python and the MySQL/python integration is great. We have a few UDFs that are home-grown but some of them were just downloaded off the net and installed. I'm sure you can find far more for MySQL than for Postgres.
There are some features in postgres that I really wish I had in MySQL: proper commit and rollback, nice/easy triggers rather than the crude enum. Support for R-trees with more than 2 dimensions (but then postgres' 3 dimensions is only a small improvement). Overall though, MySQL just seems to work where postgres can be made to work...
And, if you'd bothered to read TFA, you'd find Google complained about it more than a year before the release of Vista but was told repeatedly 'that isn't how the final version will be'.
Interesting. Any chance this will be available for users (i.e. on the iPhone?) I thought triangulating based on GSM signals was only available to carriers.
Which is funny, because I only want it to have 2GB... Since 8GB isn't enough to fit a decent percentage of my music collection on, I am going to manually manage what is on my phone (favourites + random selection + a few audio books). An extra gig of five won't be enough to change that so is not valuable to me. Now, if it went up to 80GB that would be a different matter because then I'd be able to use it differently.
Oh, and I do wish it had GPS (for the google maps integration).
3G, I dunno... my wife's phone has 3G and I have free bidirectional calling with my wife so I guess I'd get some use out of 3G if it had it, but I'm not upset by its absence.
I don't know what the accuracy rating in/. polls is, or if it is a function of the poll, so I don't know how much salt you'll have to take those results with.
The macbook has a good keyboard. Strangely the macbook pro doesn't.
The macbook pro has a good screen but the macbook doesn't (though, as noted here yesterday, the pro's is not good enough for people who object to temporal dithering). According to comments posted yesterday, Toshiba makes some laptops with screens for such people.
A good video card is subjective - I consider my 6150 a great video card because it allows me to run a 3D GUI without overheating the case. My aunt has a laptop with a built in 6600 which she got when the 6800 was the best video card available in a desktop, so gamers laptops are available. (And no, she doesn't play games, I have no idea why she got a 6600).
There are four possabilities. CEO can code, and CEO does code. I'll ignore can't and does since I've never seen it in anybody slightly senior. Then you've got can and does which, as you note, leads to micromanagement. I recently had a job where the person writing the job had three managers at different tiers saying similar but different things to them. The higher up the mananagers had more real-world experience and a better understanding of the client, but since they were busier had spent less time understanding the problem being solved.
That gives CEOs that don't program but possibly can. In my experience the ones that can't end up making the wrong decisions too often because they don't have a proper grasp of the alternatives. All it takes is a particularly slick salesman and the common-sense advice the sysadmin provided the CEO gets forgotten. Of course, it depends on your industry - outside IT I would expect these IT decisions to be sufficiently rare that there are many other skills I'd rather have my CEO know than programming. It also comes down to the CEO's ability to listen to the right people, if they have that sorted then again there is little benefit in having them harder to bamboozle.
The argument can be generalised. "It is valuable to have your CEO know how to solve the kinds of problems that your company deals with regularly". There is a school of thought that says 'CEOs are managers, they don't need to understand what the company does', and essentially I am claiming that school of thought is wrong because having the person making decisions when they don't fully understand the consequences sounds like a recipe for occasional spectacular stuffups to me.
I also disagree with you, I think you missed another requirement. Not dark ages, and not brand spanking new either.
For instance, I bought a machine based on the nvidia 6150 embedded graphics card. It was an absolute nightmare with incomatible, buggy ACPI, the network only worked after a warm boot, sound was horribly distorted with a background whine and sleep only worked once.
That nightmare lasted about three months. By then the kernel developers (with not even a bug report from me) had fixed absolutely everything wrong with it. In another three months, the fixed kernel had made its way into ubuntu and so following your advice would've worked perfectly.
For another example, what was that video card we heard about on/. the other day? The x800 or somesuch? I bet it doesn't work with the ubuntu installer... yet...
PS: There is also 'terribly made', I find really crap hardware works worse in linux than in windows. For instance, those extremely cheap quickcam thingies which require a software driver to mask gross deficencies in their sensor.
AC is mostly right. I have had 945G not up to the task in some situations, but my onboard 6200 has been more than adequate in all situations.
So, If you don't mind the odd thing going slow then 945G is fine, otherwise try to get a MB with a 6200 onboard.
If you can't do either, then there are quite a few passive 7300s around - they'll be overkill but that's your fault for picking an intel CPU but not getting the 945G:-)
Not any more. Compiz and Beryl are becoming the standard way of drawing onscreen in much the same way as aero and quartz. That means unless you have decent 3D you will stuff up desktop performance. Gamers might have much higher demands, but the days of 2D chips being adequate for desktop use are over.
(Sidetrack). There is an absolutely magnificient program for doing this that is used in the development of the GPL.
It is for text that will have every letter picked over by hundreds of different people, and I don't think it is the right tool for two people producing a document. Link:
I'm not sure you're right. Well, you are right about it being designed against nuclear attack, but I don't remember anything in DARPA being specific to smaller than nuclear?
Everything is based around: "oops, that route is down, lets try another". That sure doesn't sound scale specific to me. I would hazard a guess that either: a) Internet2 isn't designed along the same goals or b) cost cutting has lead to a structure far more similar to a tree than a graph and nowadays a well targetted bomb could take out most of the internet.
yep, and there is a port of lisp to be a shell in unix too but it doesn't quite work as a shell replacement. Could you imagine using smalltalk as your shell?
" If both from-file and to-file are directories, diff compares corre-
sponding files in both directories, in alphabetical order; this compar-
ison is not recursive unless the -r or --recursive option is given.
diff never compares the actual contents of a directory as if it were a
file. The file that is fully specified may not be standard input,
because standard input is nameless and the notion of ''file with the
same name'' does not apply. "
Amazing - I've been using diff for hmm, over a decade now, and I don't remember the last time I read the man page... Like I noticed a couple weeks ago that, since the last time I've read its man page, rsync has added a 'search for similar files with similar but different names option.
Actually, PowerShell is pretty damn good. I would say it is ahead of bash, not behind. I'm hoping somebody will steal all the good ideas and turn it into a good shell for linux.
Examples: Bash is pretty poor where your commands take more than one file and, to a lesser extent, where they produce more than one file. For instance if you have two directories a and b and want to do a diff on their contents, what do you do: % mkfifo/tmp/lsa % mkfifo/tmp/lsb % ls a >/tmp/lsa & % ls b >/tmp/lsb & % diff/tmp/ls{a,b}
That's just disgusting.
And control of GUI applications - KDE is a quite nice scripting langauge but it isn't a shell. Bash has virtually no gui scripting and I'd call its tie in with KDE's scripting poor - only the most basic datatypes can be exchanged.
Compare to applescript for an example of a fairly powerful modern scripting system.
Imagine, a whole post on the deficency of the unix shells and I didn't even mention my pet hate of escaping!
Color lasers can be had for as little as $200. I was looking at a $500 one this morning that does duplexing, PCL level 6, 30ppm, 17ppm color, ethernet (DHCP client), print spooling and...
Yes, it is more than a B/W inkjet, but it does not take long at all to justify the price difference.
At the software level, I believe the PC used to be defined as the MAC, but more recently changed to be the MB. However, Microsoft is extremely relaxed about waiving any definitions if you just give them a ring - the basic logic is that pirates don't ring the helpline so the software's tighter definition doesn't cause any problems for real customers.
"This copy of windows has been registered to another computer so cannot be registered to this computer. If you beleive this is an error, just ring 1800... now" - five minutes later you'll have a new registration code that works with the new computer and many apologies from the friendly microsoft operator for the waste of your time.
Slashdot Mirror
It is an interesting question but I disagree with you.
... well, they're mainly to stop small players and patent trolls - I think weakening them is not such a bad idea.
The judge has said that to have a patent being enforced, you must be showing goodwill efforts towards bringing a patent to market. That doesn't cut out small startups, all it cuts out (apart from patent trolls) are defensive patent portfolios and small research houses which attempt to develop and then sell ideas. I think they will be able to adapt to this - show they've sold other patents perhaps. Defensive patent portfolios
I was very interested in how well supported the hardware is or whether dell uses binary blobs. Here we are slowly shifting away from windows to ubuntu on the desktop and we already use Fedora on servers - all on dell hardware - so it will be very helpful if it is well supported hardware.
Is that a troll? Would help if you back up the assertion with something (more than IMHO). Personally I don't care who gets the deal as long as it results in roughly standard-compliant comptuers - e.g. we see more of Dell pushing ATI for decent drivers. More support for decent BIOSs, better ACPI.. there are heaps of areas where a large commodity player could help simply by offering more sales.
:-P
Oh, and to copy your trollish stance, I hope Ubuntu gets the deal
I deal with a different problem (replace web by database), but my solution might be useful to you.
I feed all logs through a log analyser (implemented using crm114.sf.net). Then the log itself is discarded and only summary statistics are kept. That way you can keep track of both longer term trends and detect when something has gone wrong.
It's a bugger to set up, but if you don't change systems that often is probably worth it.
No, nothing like D&D. It is battle reenactment except it also has fantasy sets (hence orcs) and futuristic sets (robots). There isn't any role-playing or characters - the players are sorta army generals trying to coordinate their armies.
No it isn't, Oracle is godawful to use.
It has the most awkward, painful table storage format.
ProC just sucks
Installation will leave you tearing your hair out.
Configuration is a black art.
There are many good things that can be said about Oracle - the use of indicies inside an index so you don't have to create really unnatural composite indicies... tidy SQL syntax... support for R trees of any number of dimensions... Effective subqurery optimisations... Triggers that work (compared to MySQL, not postgres). Ease of use is nowhere near the list.
Oracle is the main reason the role of DBA exists.
Firstly, lets compare:i ts_realization_at_C_for_PostgreSQL/
http://www.pgsql.cz/index.php/Project_of_UDF_and_
http://empyrean.lib.ndsu.nodak.edu/~nem/mysql/udf
AFAICT, they're both about as ugly as each other for writing your own UDFs, my point was I can download a few hundred UDFs for MySQL but I don't seem to be able to do the same thing for Postgress.
--
But say I wanted to give postgres another whirl (mostly to see if it is faster than MySQL at what I normally do). I've found SQL Fairy, is that the best approach?
I don't tend to write SQL directly. I write python code that calls the company database wrapper, which calls MySQLdb, which calls DBI, which calls MySQL. Presumably shifting to postgres involves changing the wrapper to call PyPgSQL. Is that any good, or am I in for a world of pain? I use server side cursors extensively and query the tables, but otherwise most code is just passed straight to the interpreter.
PS: I don't suppose you know if R trees are being extended in postgres beyond three dimensions? If so that would be enough justification to make me want to put serious effort in.
Hi Pavera,
I'm in a similar situation except I use MySQL. Lets look at why.
1) Ease of use: MySQL has readline, and supports lazier SQL programming. It has loads of syntactic sugar that lets you write simple things in whatever way makes sense to you. Adding users, changing passwords, backing up... it is all _easy_. Of course, postgres can do everything MySQL can too, the question is which is easier.
2) Extensive developer community. We use python and the MySQL/python integration is great. We have a few UDFs that are home-grown but some of them were just downloaded off the net and installed. I'm sure you can find far more for MySQL than for Postgres.
There are some features in postgres that I really wish I had in MySQL: proper commit and rollback, nice/easy triggers rather than the crude enum. Support for R-trees with more than 2 dimensions (but then postgres' 3 dimensions is only a small improvement). Overall though, MySQL just seems to work where postgres can be made to work...
And, if you'd bothered to read TFA, you'd find Google complained about it more than a year before the release of Vista but was told repeatedly 'that isn't how the final version will be'.
Interesting. Any chance this will be available for users (i.e. on the iPhone?)
I thought triangulating based on GSM signals was only available to carriers.
Which is funny, because I only want it to have 2GB... Since 8GB isn't enough to fit a decent percentage of my music collection on, I am going to manually manage what is on my phone (favourites + random selection + a few audio books). An extra gig of five won't be enough to change that so is not valuable to me. Now, if it went up to 80GB that would be a different matter because then I'd be able to use it differently.
Oh, and I do wish it had GPS (for the google maps integration).
3G, I dunno... my wife's phone has 3G and I have free bidirectional calling with my wife so I guess I'd get some use out of 3G if it had it, but I'm not upset by its absence.
http://slashdot.org/pollBooth.pl?qid=421&aid=-1
/. polls is, or if it is a function of the poll, so I don't know how much salt you'll have to take those results with.
http://slashdot.org/pollBooth.pl?qid=406&aid=-1
I don't know what the accuracy rating in
The macbook has a good keyboard. Strangely the macbook pro doesn't.
The macbook pro has a good screen but the macbook doesn't (though, as noted here yesterday, the pro's is not good enough for people who object to temporal dithering). According to comments posted yesterday, Toshiba makes some laptops with screens for such people.
A good video card is subjective - I consider my 6150 a great video card because it allows me to run a 3D GUI without overheating the case. My aunt has a laptop with a built in 6600 which she got when the 6800 was the best video card available in a desktop, so gamers laptops are available. (And no, she doesn't play games, I have no idea why she got a 6600).
There is some truth in both arguments.
There are four possabilities. CEO can code, and CEO does code. I'll ignore can't and does since I've never seen it in anybody slightly senior. Then you've got can and does which, as you note, leads to micromanagement. I recently had a job where the person writing the job had three managers at different tiers saying similar but different things to them. The higher up the mananagers had more real-world experience and a better understanding of the client, but since they were busier had spent less time understanding the problem being solved.
That gives CEOs that don't program but possibly can. In my experience the ones that can't end up making the wrong decisions too often because they don't have a proper grasp of the alternatives. All it takes is a particularly slick salesman and the common-sense advice the sysadmin provided the CEO gets forgotten. Of course, it depends on your industry - outside IT I would expect these IT decisions to be sufficiently rare that there are many other skills I'd rather have my CEO know than programming. It also comes down to the CEO's ability to listen to the right people, if they have that sorted then again there is little benefit in having them harder to bamboozle.
The argument can be generalised. "It is valuable to have your CEO know how to solve the kinds of problems that your company deals with regularly". There is a school of thought that says 'CEOs are managers, they don't need to understand what the company does', and essentially I am claiming that school of thought is wrong because having the person making decisions when they don't fully understand the consequences sounds like a recipe for occasional spectacular stuffups to me.
I also disagree with you, I think you missed another requirement. Not dark ages, and not brand spanking new either.
/. the other day? The x800 or somesuch? I bet it doesn't work with the ubuntu installer... yet...
For instance, I bought a machine based on the nvidia 6150 embedded graphics card. It was an absolute nightmare with incomatible, buggy ACPI, the network only worked after a warm boot, sound was horribly distorted with a background whine and sleep only worked once.
That nightmare lasted about three months. By then the kernel developers (with not even a bug report from me) had fixed absolutely everything wrong with it. In another three months, the fixed kernel had made its way into ubuntu and so following your advice would've worked perfectly.
For another example, what was that video card we heard about on
PS: There is also 'terribly made', I find really crap hardware works worse in linux than in windows. For instance, those extremely cheap quickcam thingies which require a software driver to mask gross deficencies in their sensor.
AC is mostly right. I have had 945G not up to the task in some situations, but my onboard 6200 has been more than adequate in all situations.
:-)
So, If you don't mind the odd thing going slow then 945G is fine, otherwise try to get a MB with a 6200 onboard.
If you can't do either, then there are quite a few passive 7300s around - they'll be overkill but that's your fault for picking an intel CPU but not getting the 945G
Not any more. Compiz and Beryl are becoming the standard way of drawing onscreen in much the same way as aero and quartz. That means unless you have decent 3D you will stuff up desktop performance. Gamers might have much higher demands, but the days of 2D chips being adequate for desktop use are over.
(Sidetrack). There is an absolutely magnificient program for doing this that is used in the development of the GPL.
i me_editor
It is for text that will have every letter picked over by hundreds of different people, and I don't think it is the right tool for two people producing a document. Link:
http://en.wikipedia.org/wiki/Collaborative_real-t
I'm not sure you're right. Well, you are right about it being designed against nuclear attack, but I don't remember anything in DARPA being specific to smaller than nuclear?
Everything is based around: "oops, that route is down, lets try another". That sure doesn't sound scale specific to me.
I would hazard a guess that either: a) Internet2 isn't designed along the same goals or b) cost cutting has lead to a structure far more similar to a tree than a graph and nowadays a well targetted bomb could take out most of the internet.
yep, and there is a port of lisp to be a shell in unix too but it doesn't quite work as a shell replacement.
Could you imagine using smalltalk as your shell?
Cool! I'm certainly going to start using that :)
" If both from-file and to-file are directories, diff compares corre-
sponding files in both directories, in alphabetical order; this compar-
ison is not recursive unless the -r or --recursive option is given.
diff never compares the actual contents of a directory as if it were a
file. The file that is fully specified may not be standard input,
because standard input is nameless and the notion of ''file with the
same name'' does not apply.
"
Amazing - I've been using diff for hmm, over a decade now, and I don't remember the last time I read the man page...
Like I noticed a couple weeks ago that, since the last time I've read its man page, rsync has added a 'search for similar files with similar but different names option.
Actually, PowerShell is pretty damn good. I would say it is ahead of bash, not behind. I'm hoping somebody will steal all the good ideas and turn it into a good shell for linux.
/tmp/lsa /tmp/lsb /tmp/lsa & /tmp/lsb & /tmp/ls{a,b}
Examples: Bash is pretty poor where your commands take more than one file and, to a lesser extent, where they produce more than one file.
For instance if you have two directories a and b and want to do a diff on their contents, what do you do:
% mkfifo
% mkfifo
% ls a >
% ls b >
% diff
That's just disgusting.
And control of GUI applications - KDE is a quite nice scripting langauge but it isn't a shell. Bash has virtually no gui scripting and I'd call its tie in with KDE's scripting poor - only the most basic datatypes can be exchanged.
Compare to applescript for an example of a fairly powerful modern scripting system.
Imagine, a whole post on the deficency of the unix shells and I didn't even mention my pet hate of escaping!
Windows is extremely cheap in mobile devices. It has downsides - but cost per unit is usually not at the top of the list.
Color lasers can be had for as little as $200. I was looking at a $500 one this morning that does duplexing, PCL level 6, 30ppm, 17ppm color, ethernet (DHCP client), print spooling and ...
Yes, it is more than a B/W inkjet, but it does not take long at all to justify the price difference.
At the software level, I believe the PC used to be defined as the MAC, but more recently changed to be the MB. However, Microsoft is extremely relaxed about waiving any definitions if you just give them a ring - the basic logic is that pirates don't ring the helpline so the software's tighter definition doesn't cause any problems for real customers.
... now" - five minutes later you'll have a new registration code that works with the new computer and many apologies from the friendly microsoft operator for the waste of your time.
"This copy of windows has been registered to another computer so cannot be registered to this computer. If you beleive this is an error, just ring 1800