That's bull. H.264 is the standard these days, except for pirate content or patent encumbered formats like MS' video formats. I can rip my DVD's into H264 quite handily without noticeable loss of quality.
I think you missed Trance's point. Say 5 years ago you had gone through your DVD library and carefully ripped everything to DivX. Perhaps you lived in a country where it's legal to rip DVDs you got out from the video store... Regardless, you can easily have built up a decent collection of DivX encoded movies. You can say H.264 is the standard these days, and you'd be right, but I only changed my default when Handbrake changed its default and I imagine I'm pretty typical. I don't keep up with the changes of video codecs well enough to know when the default is not the best choice.
Converting this video library to H.264 would cost quality - You'd have to go back to the source DVDs to get a decent rip. It would be like if iTunes decided tomorrow to drop support for MP3 because AAC is the standard now. I personally am perfectly happy with AAC but I've thrown out my old CDs to save space years ago after ripping them to high quality MP3. There's a big difference between encouraging a modern format and preventing the use of legacy formats.
All platforms have issues,
Yes they do. But on other platforms it is possible to change their behaviour to match how you like to work. Apple has very good default behaviour but it has extremely limited support for changing that behaviour. For instance, if I use my iPhone all the time as a camera and I really want to use the volume buttons to zoom then I'm out of luck. Apple mandates what it considers best practice and as a user I am frequently unable to override this even on my computer.
Sorry to reply twice but I just thought of a different example for you.
New Zealand, early 90s... Sauvignon Blanc is the young rising star. Everyone is ripping up old Chardonnay to plant it and the fashionable set is all over it.
The problem is that palates familiar with great big chardonnay were finding this wine a bit sharp and thin. Effectively, it was selling because it had become fashionable, not because the general population liked it yet.
Wineries were faced with the problem of how do they sell something called Sauvignon Blanc that their customers will enjoy drinking. Well, one enterprising company solved it by simply lying. The details are hazy but as I recall they put about 25% umm semillion? in to soften it. Probably a good idea for palates at the time but not true to label.
A wine judge was able to detect the addition and In the resulting scandal it was discovered that NZ wine laws do not have any assurance guarantees unlike the EU. So while there was a bit of a backlash for the damage they did to the NZ brand, they largely got away with it.
Two reasons spring to mind - both advantaging the people doing it at the cost of the appellation's brand:
1) People outside the area will write Sauterne or Eiswein on any sweet wine. Buyers go away with the wrong impression of what a Sauterne is.
2) People within the area are trying to differentiate themselves from he dozens of neighboring vineyards. One way is to apply different winemaking techniques, so our wine tastes distinctly different. Ok for them, bade for the overall brand.
In both cases it is possible the impostor using the brand is a really good wine. But it is no longer DOCG or whatever standard is appropriate. It's effectively trademark infringement.
I think you've missed the point. The purpose of the names like Bordeaux, Burgundy, Chianti, etc. is not to tell you that it is good, though it does usally tell you that it is at least ok. It is to tell you that it is in the style that the area is famous for. An Australian Pinot Noir might be stunning, but you can't meaningfully call it Burgundy because it isn't that style. It might be better than every wine made in Burgundy, but it still _isn't_ burgundy.
If Australia's winemakers ever cooperate enough to develop a distinct style that's consistent along say the Barossa valley say then by all means call it Barossa wine instead of Shiraz. But until then, I think it's much clearer to talk about the quality of Australian wine and use a generic name like Chardonnay rather than the name of a region in France that probably does not stylistically match the Australian wine anyway.
Even the Europeans do this. If you are making wine in Chianti and want to do something differently then you _cannot_ call your wine Chianti - because it isn't wine made in the style of that region. What it means is that when you pick up a bottle of Chianti, you know what you're buying (though not the quality). Australian Chardonnay could be anything, from a subtle unoaked variety to a monster.
I think wave failed because it did not have a transition path.
Wave is for collaboration, it was pretty much useless for just one person. Gmail interacts with any SMTP server so it was easy to grow organically. I think wave was a similar step above gmail as gmail was to webmail at the time. However wave made no real attempt to interact with legacy systems (even google legacy systems like gmail, google talk, google docs) and so with wave it was almost like joining a gated community and it quickly got boring...
In simple terms it will make things faster. Beyond that it's hard to say.
SSDs have fundamentally different performance characteristics. Their write speed is asymmetrical so I imagine they'll be more popular in data warehousing than in operational databases, and their complete lack of seek time means that some aspects of database engine design (e.g. b-trees) are fundamentally no longer relevant. At a guess this means that hash join indices will be much more realistic because it becomes possible to simultaneously access many indices without a performance hit. With any luck that will push column centric data storage even further as people realise they can index every column and get far greater flexibility.
However I don't think changes will come very quickly. Any really high performance database is already in RAM and moderately high performance databases have so many spindles that they're getting a similar sort of effect already.
That's interesting, I have not come across this before.
I last worked in computational linguistics over five years ago and but when I left there were a good supply of techniques for automatically extracting meaning from an unknown text.
My own research was able to build up both a dendrogram and word vectors from any sufficiently large corpus, and a quick google search turned up http://www.springerlink.com/content/fp17278783422256/ which shows that the field is continuing to develop. I would expect that by now it would be pretty easy to feed a text like this in and get word associations out. From your word associations, building up a basic dictionary will still need you to bootstrap associated concepts but at least the task is much smaller and there's a lot of support for checking.
I don't recall much successful research into automatic parsing of unknown languages, but since I left the field it could've progressed. Shallow parsing would be a good place to start. Since the language's stemming is unknown you're going to be hard-pressed to parse it anyway but POS tagging should be doable.
I have not done any work with cyphered texts, so I'm assuming that approaches to natural languages will apply. No doubt there is research in this area, I'm just not familiar with it.
I've had a production Oracle database go down due to a bug in Oracle and raised a support ticket. It took Oracle a week to produce a workaround - what kind of environment can afford to have the production database down for a week? I consider Oracle support okay if you're doing development and don't require quick turnaround but pretty much worthless for handling P1 incidents.
I didn't. I got burned too often by small projects having to be changed because of licencing issues.
(rant starting) For instance I had a 'team' (of 2) doing advanced analytics that were hammering the data warehouse. It was proposed to use RAC and have them access another machine - for the cost of RAC I could buy a whole team to write custom replication code! Then it was proposed to replicate using some simple ETL code to another database, canned due to the cost of another production Oracle licence just to satisfy two users. In the end this is pretty much what I went with, except that the destination machine was running SAS because it's cheaper and the advanced analytic users prefer it.
Same company, different project... the data warehouse was being asked by the business to perform a number of fixes to raw source data that had the DW manager cringing. He proposed splitting the DW into a proper DW and put another machine in between operations and the DW which performed these fixes. The project had to be canned because we couldn't justify the extra Oracle licences just for process purity.
The conclusion I took from this is that Oracle may be better than MS SQL or Postgres and it may be better in ways that justify the extra cost (locking becomes almost a non-issue, etc.) but if you go with Oracle then you'll end up preventing all those cool little improvements that your staff think of over the years and I'm not willing to wear that level of inflexibility.
So the last time I had to make this call I went with MS SQL. Yes I've had issues that I wouldn't have had with Oracle, and I got a lot of flack from the Oracle sales rep about the call, but overall it's performed pretty well. One of the things the rep pointed out was that Oracle standard edition was cheaper than MS standard and contained support for RAC (which would've been nice) but I kept remembering how many Oracle things are only in EE and imagining the licencing costs in say five years time when my successor finds they have no choice except to migrate to EE and looks at the use of RAC I've made. I don't want to hobble the future flexibility of companies I work for like that.
There's quite a few things that MS does better than Oracle too, just they tend to not in the core DB engine. Reporting services is pretty slick, Integration Services versus OWB..., Management Studio is not bad either, MS SQL's spatial might be less featureful than Oracle Spatial but it's ten times easier to code for, and MS SQL's integration with CLR makes up for much of PL/SQL's superiority over T-SQL (which sucks).
Next time I hope to try Postgres, since an annoying side effect of choosing MS SQL is the server has to run Windows, but we'll see. Unless I can be very sure the system architecture won't change, Oracle will be about as likely to be chosen as MySQL - basically no unless my staff only have experience with it. Postgres' support for embedded languages is better than Oracle's too...
PDF is an open format and should load in 25 years just just fine. Technically, it's practically just a zipped Postscript which was invented about 25 years ago. DOC I suggest you save to PDF.
There's lots of advantages to XML over PDF - metadata support for one - reliable plain text extraction for another - but it is technically an open standard.
I had similar views to yours. Then I happened to get a summer job working for a desktop publisher and so had to use photoshop. I won't claim to be an expert - and I'm awfully rusty now - but you can do very awesome things in Photoshop extremely easily if you happen to have spent a large number of hours learning how. Yes, Photoshop is hard to _learn_ but it is very easy to use.
You say that ordinary users just need to , adjust brightness etc.but I don't think this is true. Ordinary users want to tune up their photos - e.g. sharpen, remove the shadow from someone's face, take the reflection off someone's glasses, remove a lamp-post or cyclist that unfortunately interfere with the shot, replace the blinking eyes from one photo with the open eyes from the next (especially group photos where someone is invariably looking away), etc, slightly fancier resize (e.g. fix camera not quite straight).
Also, my bet is that my list of basic features and the guy next to me's list will not be identical - if you want to make all basic users happy then I suspect you'll be in for a big list of features. For instance a grandmother with a thousand old photos in a shoe box will have a very different basic list to the one I gave above involving scratch removal and the like.
Now, I've completely avoided answering your question. Instead I've told you to invest the time in learning gimp, it will pay off over the years. In terms of actually answering your question I haven't found a good answer - Apple's Aperture is an attempt, and Adobe makes Photoshop Elements but they all suck
I'm tempted to say RTFA but in the interest of saving you and no doubt others a bit of time:
"The biggest mistake Apple did with FileVault is storing the encrypted home directory on a virtual file system. All of FileVault's drawbacks originate from this. The implementation is brilliant, free of bugs, fast and well thought over. But why they decided to have all the trouble with a filesystem in a filesystem remains a mystery."
Essentially, instead of mounting/Users/your_username via FIleVault, Apple decided to add a sparse bundle file to your home directory with all of the contents. The worst impact of this design flaw is it adds a lot of time overhead at log out. If apple instead created a different partition for each user's home directory then there are no real flaws with FileVault.
I can see why Apple did it they way they did - dynamically resizing partitions as the user adds data to their home directory sounds... scary.
Perhaps it will cost a teacher - though I think more likely you'll lose something optional - like a librarian. But 50 busses * 30 kids * 2 hour long trips a day is 3000 hours of benefit every day. The librarian would no doubt have benefitted a few kids far more, but I'd be that on average this is better.
I have a hard time imagining how a school could get a similar amount of benefit for $800/year.
I'm thinking of buying one so I guess I'm qualified to answer.
I'm viewing it as an entertainment device. Compared to my iphone it is: * Less portable * Faster * Much bigger screen * Better resolution * Possibly other things I don't know/care about
Now I'd say most people out there like being entertained so that sounds to me like it will appeal to a lot of people. The things you've asked for, like Office software... well I don't know what kind of nerd you are but I don't find Office software at all entertaining.
Slashdot Mirror
Er... yes it would.
That's bull. H.264 is the standard these days, except for pirate content or patent encumbered formats like MS' video formats. I can rip my DVD's into H264 quite handily without noticeable loss of quality.
I think you missed Trance's point. Say 5 years ago you had gone through your DVD library and carefully ripped everything to DivX. Perhaps you lived in a country where it's legal to rip DVDs you got out from the video store... Regardless, you can easily have built up a decent collection of DivX encoded movies. You can say H.264 is the standard these days, and you'd be right, but I only changed my default when Handbrake changed its default and I imagine I'm pretty typical. I don't keep up with the changes of video codecs well enough to know when the default is not the best choice.
Converting this video library to H.264 would cost quality - You'd have to go back to the source DVDs to get a decent rip. It would be like if iTunes decided tomorrow to drop support for MP3 because AAC is the standard now. I personally am perfectly happy with AAC but I've thrown out my old CDs to save space years ago after ripping them to high quality MP3. There's a big difference between encouraging a modern format and preventing the use of legacy formats.
All platforms have issues,
Yes they do. But on other platforms it is possible to change their behaviour to match how you like to work. Apple has very good default behaviour but it has extremely limited support for changing that behaviour. For instance, if I use my iPhone all the time as a camera and I really want to use the volume buttons to zoom then I'm out of luck. Apple mandates what it considers best practice and as a user I am frequently unable to override this even on my computer.
No, the iPhone has quite limited enterprise features - the ability to mass deploy changes and lock them down tight is all but non-existent
That's very clever... I think you could get that to catch on
Sorry to reply twice but I just thought of a different example for you.
New Zealand, early 90s... Sauvignon Blanc is the young rising star. Everyone is ripping up old Chardonnay to plant it and the fashionable set is all over it.
The problem is that palates familiar with great big chardonnay were finding this wine a bit sharp and thin. Effectively, it was selling because it had become fashionable, not because the general population liked it yet.
Wineries were faced with the problem of how do they sell something called Sauvignon Blanc that their customers will enjoy drinking. Well, one enterprising company solved it by simply lying. The details are hazy but as I recall they put about 25% umm semillion? in to soften it. Probably a good idea for palates at the time but not true to label.
A wine judge was able to detect the addition and In the resulting scandal it was discovered that NZ wine laws do not have any assurance guarantees unlike the EU. So while there was a bit of a backlash for the damage they did to the NZ brand, they largely got away with it.
Two reasons spring to mind - both advantaging the people doing it at the cost of the appellation's brand:
1) People outside the area will write Sauterne or Eiswein on any sweet wine. Buyers go away with the wrong impression of what a Sauterne is.
2) People within the area are trying to differentiate themselves from he dozens of neighboring vineyards. One way is to apply different winemaking techniques, so our wine tastes distinctly different. Ok for them, bade for the overall brand.
In both cases it is possible the impostor using the brand is a really good wine. But it is no longer DOCG or whatever standard is appropriate. It's effectively trademark infringement.
I think you've missed the point. The purpose of the names like Bordeaux, Burgundy, Chianti, etc. is not to tell you that it is good, though it does usally tell you that it is at least ok. It is to tell you that it is in the style that the area is famous for. An Australian Pinot Noir might be stunning, but you can't meaningfully call it Burgundy because it isn't that style. It might be better than every wine made in Burgundy, but it still _isn't_ burgundy.
If Australia's winemakers ever cooperate enough to develop a distinct style that's consistent along say the Barossa valley say then by all means call it Barossa wine instead of Shiraz. But until then, I think it's much clearer to talk about the quality of Australian wine and use a generic name like Chardonnay rather than the name of a region in France that probably does not stylistically match the Australian wine anyway.
Even the Europeans do this. If you are making wine in Chianti and want to do something differently then you _cannot_ call your wine Chianti - because it isn't wine made in the style of that region. What it means is that when you pick up a bottle of Chianti, you know what you're buying (though not the quality). Australian Chardonnay could be anything, from a subtle unoaked variety to a monster.
Well, I disagree with you :)
I think wave failed because it did not have a transition path.
Wave is for collaboration, it was pretty much useless for just one person. Gmail interacts with any SMTP server so it was easy to grow organically. I think wave was a similar step above gmail as gmail was to webmail at the time. However wave made no real attempt to interact with legacy systems (even google legacy systems like gmail, google talk, google docs) and so with wave it was almost like joining a gated community and it quickly got boring...
In simple terms it will make things faster. Beyond that it's hard to say.
SSDs have fundamentally different performance characteristics. Their write speed is asymmetrical so I imagine they'll be more popular in data warehousing than in operational databases, and their complete lack of seek time means that some aspects of database engine design (e.g. b-trees) are fundamentally no longer relevant. At a guess this means that hash join indices will be much more realistic because it becomes possible to simultaneously access many indices without a performance hit. With any luck that will push column centric data storage even further as people realise they can index every column and get far greater flexibility.
However I don't think changes will come very quickly. Any really high performance database is already in RAM and moderately high performance databases have so many spindles that they're getting a similar sort of effect already.
Well, it would be pretty handy if I could format an external drive in something other than FAT32 or NTFS for support on different systems...
Is redhat an open source company? I see no difference between what Sugar is doing and RHEL - except perhaps RedHat is more open about it.
Right, from what I read it's exactly like CentOS vs RHEL. Whether that's open source or open core comes down to definitions more than anything.
Given that ACM continue to oppose software patents, that seems an easy place to start.
That's interesting, I have not come across this before.
I last worked in computational linguistics over five years ago and but when I left there were a good supply of techniques for automatically extracting meaning from an unknown text.
My own research was able to build up both a dendrogram and word vectors from any sufficiently large corpus, and a quick google search turned up http://www.springerlink.com/content/fp17278783422256/ which shows that the field is continuing to develop. I would expect that by now it would be pretty easy to feed a text like this in and get word associations out. From your word associations, building up a basic dictionary will still need you to bootstrap associated concepts but at least the task is much smaller and there's a lot of support for checking.
I don't recall much successful research into automatic parsing of unknown languages, but since I left the field it could've progressed. Shallow parsing would be a good place to start. Since the language's stemming is unknown you're going to be hard-pressed to parse it anyway but POS tagging should be doable.
I have not done any work with cyphered texts, so I'm assuming that approaches to natural languages will apply. No doubt there is research in this area, I'm just not familiar with it.
They wrote it, apple rejected it and told them to make a webapp - google for more details.
A bit hard now as I'm no longer working at that company.
In case it is relevant, I live in New Zealand where Oracle have no 'on the ground' technical staff.
You're serious?
I've had a production Oracle database go down due to a bug in Oracle and raised a support ticket. It took Oracle a week to produce a workaround - what kind of environment can afford to have the production database down for a week? I consider Oracle support okay if you're doing development and don't require quick turnaround but pretty much worthless for handling P1 incidents.
I didn't. I got burned too often by small projects having to be changed because of licencing issues.
(rant starting)
For instance I had a 'team' (of 2) doing advanced analytics that were hammering the data warehouse. It was proposed to use RAC and have them access another machine - for the cost of RAC I could buy a whole team to write custom replication code! Then it was proposed to replicate using some simple ETL code to another database, canned due to the cost of another production Oracle licence just to satisfy two users. In the end this is pretty much what I went with, except that the destination machine was running SAS because it's cheaper and the advanced analytic users prefer it.
Same company, different project... the data warehouse was being asked by the business to perform a number of fixes to raw source data that had the DW manager cringing. He proposed splitting the DW into a proper DW and put another machine in between operations and the DW which performed these fixes. The project had to be canned because we couldn't justify the extra Oracle licences just for process purity.
The conclusion I took from this is that Oracle may be better than MS SQL or Postgres and it may be better in ways that justify the extra cost (locking becomes almost a non-issue, etc.) but if you go with Oracle then you'll end up preventing all those cool little improvements that your staff think of over the years and I'm not willing to wear that level of inflexibility.
So the last time I had to make this call I went with MS SQL. Yes I've had issues that I wouldn't have had with Oracle, and I got a lot of flack from the Oracle sales rep about the call, but overall it's performed pretty well. One of the things the rep pointed out was that Oracle standard edition was cheaper than MS standard and contained support for RAC (which would've been nice) but I kept remembering how many Oracle things are only in EE and imagining the licencing costs in say five years time when my successor finds they have no choice except to migrate to EE and looks at the use of RAC I've made. I don't want to hobble the future flexibility of companies I work for like that.
There's quite a few things that MS does better than Oracle too, just they tend to not in the core DB engine. Reporting services is pretty slick, Integration Services versus OWB..., Management Studio is not bad either, MS SQL's spatial might be less featureful than Oracle Spatial but it's ten times easier to code for, and MS SQL's integration with CLR makes up for much of PL/SQL's superiority over T-SQL (which sucks).
Next time I hope to try Postgres, since an annoying side effect of choosing MS SQL is the server has to run Windows, but we'll see. Unless I can be very sure the system architecture won't change, Oracle will be about as likely to be chosen as MySQL - basically no unless my staff only have experience with it. Postgres' support for embedded languages is better than Oracle's too...
And... what's wrong with using iPhoto to burn your pictures because your hard drive is full?
Only iPhoto will know for sure what data it needs to burn to CD in an export.
PDF is an open format and should load in 25 years just just fine. Technically, it's practically just a zipped Postscript which was invented about 25 years ago.
DOC I suggest you save to PDF.
There's lots of advantages to XML over PDF - metadata support for one - reliable plain text extraction for another - but it is technically an open standard.
No, it's going fairly well from a long term perspective.
Just seems that some people object to the pain and suffering necessary for this method to work.
I had similar views to yours. Then I happened to get a summer job working for a desktop publisher and so had to use photoshop. I won't claim to be an expert - and I'm awfully rusty now - but you can do very awesome things in Photoshop extremely easily if you happen to have spent a large number of hours learning how. Yes, Photoshop is hard to _learn_ but it is very easy to use.
You say that ordinary users just need to , adjust brightness etc.but I don't think this is true. Ordinary users want to tune up their photos - e.g. sharpen, remove the shadow from someone's face, take the reflection off someone's glasses, remove a lamp-post or cyclist that unfortunately interfere with the shot, replace the blinking eyes from one photo with the open eyes from the next (especially group photos where someone is invariably looking away), etc, slightly fancier resize (e.g. fix camera not quite straight).
Also, my bet is that my list of basic features and the guy next to me's list will not be identical - if you want to make all basic users happy then I suspect you'll be in for a big list of features. For instance a grandmother with a thousand old photos in a shoe box will have a very different basic list to the one I gave above involving scratch removal and the like.
Now, I've completely avoided answering your question. Instead I've told you to invest the time in learning gimp, it will pay off over the years. In terms of actually answering your question I haven't found a good answer - Apple's Aperture is an attempt, and Adobe makes Photoshop Elements but they all suck
I'm tempted to say RTFA but in the interest of saving you and no doubt others a bit of time:
"The biggest mistake Apple did with FileVault is storing the encrypted home directory on a virtual file system. All of FileVault's drawbacks originate from this. The implementation is brilliant, free of bugs, fast and well thought over. But why they decided to have all the trouble with a filesystem in a filesystem remains a mystery."
Essentially, instead of mounting /Users/your_username via FIleVault, Apple decided to add a sparse bundle file to your home directory with all of the contents. The worst impact of this design flaw is it adds a lot of time overhead at log out. If apple instead created a different partition for each user's home directory then there are no real flaws with FileVault.
I can see why Apple did it they way they did - dynamically resizing partitions as the user adds data to their home directory sounds... scary.
Perhaps it will cost a teacher - though I think more likely you'll lose something optional - like a librarian. But 50 busses * 30 kids * 2 hour long trips a day is 3000 hours of benefit every day. The librarian would no doubt have benefitted a few kids far more, but I'd be that on average this is better.
I have a hard time imagining how a school could get a similar amount of benefit for $800/year.
I'm thinking of buying one so I guess I'm qualified to answer.
I'm viewing it as an entertainment device. Compared to my iphone it is:
* Less portable
* Faster
* Much bigger screen
* Better resolution
* Possibly other things I don't know/care about
Now I'd say most people out there like being entertained so that sounds to me like it will appeal to a lot of people. The things you've asked for, like Office software... well I don't know what kind of nerd you are but I don't find Office software at all entertaining.