But the dev time for tables is often much shorter, especially if you don't deal with CSS as your day job. This isn't just on IE either I mean the "rules" off CSS:P are just bizzare and you end up having to nest divs with float params and negative margins left right and centre resulting in almost as much complexity as the tables your were trying to avoid.
I dont think css (at least in its current incarnation) or tables is the right solution and hope we see some innovation in this area.
Thats really interesting. Do you remember the name of the study? I would like to read it. Did it go as far as to extrapolate what standard of life we would get if we evenly distributed the combined wealth of just one Earth?
Great post and you've perfectly highlighted why the fast proportin of open source software has a tiny market share even when the marginal cost is zero.
In the real world companies which succeed are ones which either reduce the costs of their customers or make their customers more money. If your customer isnt reducing costs because they cant use your software effectively then even if its free the TCO is much higher than commercial software since the customers time will very quickly become more valuable than the software.
UI is just a thin layer though. Some problems are too deep to realistically fix with a facelift. For example if your FOSS developer has designed a tool that requires you to learn some convoluted homebrewed gammar to pass in as a file then no amount of UI tinkering is going to fix it.
Similarly some projects like libraries dont have a UI (beyond a clean API I suppose). However a fundamental mistake in the architecture is unlikely to be fixable with any amount of API rejigging.
User experience needs to be implemented from the ground up which is a fundamental difference in how FOSS and commercial programmers operate. FOSS guys, at least the hobbiest kind tend to think about "Im going to write an implementation of algorithm X or tech Y". In the commercial word it starts with "We're going to solve user problem Z".
I know what your saying but even if the complainant is a coder submitting a patch still has a high barrier of entry for all but the most trivial of programs. For example lets say I discovered a flaw with how firefox implemented some obscure css property under certain conditions. How long do you think it would take me to become familiar enough with the codebase to fix it. What about if its a project where the source code isnt documented.
Alot of FOSS software doesnt have any design docs, the architecture lives in the developers head. At least when your working on a commercial project you'll have access to internal technical documentation.
So basically what Im saying is that from a cold start submitting a patch is pretty work intensive and unrealistic. Does that mean you should provide free support? Of course not but you could be more helpful by suggesting where in the codebase is a good place to start looking or giving a 10,000 feet over view of how that area of the code hangs together.
Actually when IE7 first came out it had a couple of nicer UI features. At the time for example the close button for tabs war on the far right hand side of the tab bar rather than on the individual tabs themselves. Earliest version of FF i have on me is 2.0.0.16 in which its fixed so it must have been an early 2 release.
Linux and most UNIX systems and applications are harder to use because they are built with the architecture of the code in mind. A good UNIX program can easily work with other UNIX programs, and a good UNIX program is made as general as possible to maximize speed and reduce bloat as the program advances.
The unix philosophy is actually do one thing and do it well not be as general as possible. Which is fine for simple command line utilities but in many areas users are concerned with workflow. They want to use one application or suite of applications to acheive the tasks they need to do in a particular problem domain with minimal fuss. Not string together 14 different command line tools in an arse backwards shell scripting language.
Well then to be honest your task is probably either not very complicated or is one which the command line happens to support very well like programming. I'd argue that even for coding a GUI is much more useful in many cases, theres a myriad of code inspection based help, autocomplete, call hierarchies and try using dbx from the comand line compared to the visual studio or eclipse java debuggers.
I dont find word and latex compete. In my thesis I used both. I wrote the main chunk in word for ease, spelling / grammar checker etc. then copy / pasted into a latex document doing the refs as I went.
It depends what you're looking at. If you're looking at a system of geographically dispersed machines then in a correctly designed app a power outage shouldnt lead to any data loss. A transactional model puts a strict precondition on participating resources asserting that they must be able to commit or roll back, this effectively means that everything is logged before it is committed meaning a power outage at any stage in the transaction is recoverable.
So presumably JimBob@openid.google.com is a different person from JimBob@mymaliciousopenidprovider.ru from the site's perspective?
Which means your essentially tying yourself to one openID provider?
Or the commuter could use tokens which bear a relationship to the fare. They could obtain these tokens by putting their card into handily located dispensing machines which would then debit the equivelent amount of tokens from their account.
If only such a system existed.
The mass of boring, specific-solution apps out there dwwarf everything you can get commercially. Windows is built on the premise that it is easy to create apps, and that supporting them is easy even if the original developer leave, you'll be able to find another who can take up their code because they will be familiar with the technology used to produce it.
I think you're onto something there. Do one thing and do it well is a great way to build reusable code packages but isn't a good workflow. The idea of chaining output thorugh command line utilities and writing shell scripts to acheive what should be workflow options in some kind of domain specific graphiucal suite is long over for all but developers.
Did your paranoia contribute towards your divorce? Cos most women I know would certainly be turned off by frothing diatribes about "lojacking" kids asses in aminute with no choice.
Because they need a Visual Studio environment as you would know if you had read the summary..I know reading the article is a stretch but really this is ridiculous.
When you see the designed for XP or vista stickers on the copmuters at work to you feel the need to write "or linux" on the bottom?
The problems the spammers are solving aren't really very interesting and mostly hinge on repeating something millions of times. It might interest you to know that the state of some elements of AI is actually very advanced for example free text can be correctly POS tagged with 95% accuracy in French (I beleive English is slightly lower). OCR hasnt seen active research in decades and object classification and feature recognition are much more active research topics.
Cry me a bleeding river. The number of times I've been sniped by a power seller who then goes on to resell the item is ridiculous. Ebay shouldnt be an alternative to getting a job. And they certainly dont have an obligation to support the business models of these parasitic individuals. I liked it when it was casual auction site. And it hasnt been that way for a LOOOONG time.
The OS itself is pretty mature, theres still the case of having to edit config files that arises on occasion but the real problem tying people to windows is applications. If your doing any task that requires workflow you generally find a very good proprietary product on Windows and an immature variant on linux. For me for example the windows killer product is cubase. There is just nothing remotely close available for linux. For others that could be photoshop or maya or whatever.
For business they have their custom application which may have been running since the 3.11 days and are still soldiering away under XP. As I said its applications that are holding back the switch.
Slashdot Mirror
But the dev time for tables is often much shorter, especially if you don't deal with CSS as your day job. This isn't just on IE either I mean the "rules" off CSS:P are just bizzare and you end up having to nest divs with float params and negative margins left right and centre resulting in almost as much complexity as the tables your were trying to avoid.
I dont think css (at least in its current incarnation) or tables is the right solution and hope we see some innovation in this area.
Thats really interesting. Do you remember the name of the study? I would like to read it. Did it go as far as to extrapolate what standard of life we would get if we evenly distributed the combined wealth of just one Earth?
Great post and you've perfectly highlighted why the fast proportin of open source software has a tiny market share even when the marginal cost is zero.
In the real world companies which succeed are ones which either reduce the costs of their customers or make their customers more money. If your customer isnt reducing costs because they cant use your software effectively then even if its free the TCO is much higher than commercial software since the customers time will very quickly become more valuable than the software.
UI is just a thin layer though. Some problems are too deep to realistically fix with a facelift. For example if your FOSS developer has designed a tool that requires you to learn some convoluted homebrewed gammar to pass in as a file then no amount of UI tinkering is going to fix it.
Similarly some projects like libraries dont have a UI (beyond a clean API I suppose). However a fundamental mistake in the architecture is unlikely to be fixable with any amount of API rejigging.
User experience needs to be implemented from the ground up which is a fundamental difference in how FOSS and commercial programmers operate. FOSS guys, at least the hobbiest kind tend to think about "Im going to write an implementation of algorithm X or tech Y". In the commercial word it starts with "We're going to solve user problem Z".
I know what your saying but even if the complainant is a coder submitting a patch still has a high barrier of entry for all but the most trivial of programs. For example lets say I discovered a flaw with how firefox implemented some obscure css property under certain conditions. How long do you think it would take me to become familiar enough with the codebase to fix it. What about if its a project where the source code isnt documented.
Alot of FOSS software doesnt have any design docs, the architecture lives in the developers head. At least when your working on a commercial project you'll have access to internal technical documentation.
So basically what Im saying is that from a cold start submitting a patch is pretty work intensive and unrealistic. Does that mean you should provide free support? Of course not but you could be more helpful by suggesting where in the codebase is a good place to start looking or giving a 10,000 feet over view of how that area of the code hangs together.
Actually when IE7 first came out it had a couple of nicer UI features. At the time for example the close button for tabs war on the far right hand side of the tab bar rather than on the individual tabs themselves. Earliest version of FF i have on me is 2.0.0.16 in which its fixed so it must have been an early 2 release.
Linux and most UNIX systems and applications are harder to use because they are built with the architecture of the code in mind. A good UNIX program can easily work with other UNIX programs, and a good UNIX program is made as general as possible to maximize speed and reduce bloat as the program advances.
The unix philosophy is actually do one thing and do it well not be as general as possible. Which is fine for simple command line utilities but in many areas users are concerned with workflow. They want to use one application or suite of applications to acheive the tasks they need to do in a particular problem domain with minimal fuss. Not string together 14 different command line tools in an arse backwards shell scripting language.
Well then to be honest your task is probably either not very complicated or is one which the command line happens to support very well like programming. I'd argue that even for coding a GUI is much more useful in many cases, theres a myriad of code inspection based help, autocomplete, call hierarchies and try using dbx from the comand line compared to the visual studio or eclipse java debuggers.
I dont find word and latex compete. In my thesis I used both. I wrote the main chunk in word for ease, spelling / grammar checker etc. then copy / pasted into a latex document doing the refs as I went.
I'm gonna squat "Alegbra" with an advert to my domain hosting service.
My table saw is mission critical...*especially* during a power outage.
It depends what you're looking at. If you're looking at a system of geographically dispersed machines then in a correctly designed app a power outage shouldnt lead to any data loss. A transactional model puts a strict precondition on participating resources asserting that they must be able to commit or roll back, this effectively means that everything is logged before it is committed meaning a power outage at any stage in the transaction is recoverable.
So presumably JimBob@openid.google.com is a different person from JimBob@mymaliciousopenidprovider.ru from the site's perspective? Which means your essentially tying yourself to one openID provider?
Isnt that what POSIX is for? Although admitedly it says nothing about the GUI side of things.
Or the commuter could use tokens which bear a relationship to the fare. They could obtain these tokens by putting their card into handily located dispensing machines which would then debit the equivelent amount of tokens from their account.
If only such a system existed.
The mass of boring, specific-solution apps out there dwwarf everything you can get commercially. Windows is built on the premise that it is easy to create apps, and that supporting them is easy even if the original developer leave, you'll be able to find another who can take up their code because they will be familiar with the technology used to produce it.
I think you're onto something there. Do one thing and do it well is a great way to build reusable code packages but isn't a good workflow. The idea of chaining output thorugh command line utilities and writing shell scripts to acheive what should be workflow options in some kind of domain specific graphiucal suite is long over for all but developers.
I always wondered if Heisenberg's uncertaintly principle was adequate defence against speeding tickets.
Did your paranoia contribute towards your divorce? Cos most women I know would certainly be turned off by frothing diatribes about "lojacking" kids asses in aminute with no choice.
....and with less cores.
Because they need a Visual Studio environment as you would know if you had read the summary..I know reading the article is a stretch but really this is ridiculous.
When you see the designed for XP or vista stickers on the copmuters at work to you feel the need to write "or linux" on the bottom?
The problems the spammers are solving aren't really very interesting and mostly hinge on repeating something millions of times. It might interest you to know that the state of some elements of AI is actually very advanced for example free text can be correctly POS tagged with 95% accuracy in French (I beleive English is slightly lower). OCR hasnt seen active research in decades and object classification and feature recognition are much more active research topics.
Cry me a bleeding river. The number of times I've been sniped by a power seller who then goes on to resell the item is ridiculous. Ebay shouldnt be an alternative to getting a job. And they certainly dont have an obligation to support the business models of these parasitic individuals. I liked it when it was casual auction site. And it hasnt been that way for a LOOOONG time.
Unfortunately most people think its a database.
Sysadmin skills are a commodity. Highly specialised application programming skills are an asset. I think your advice is wholly incorrect.
The OS itself is pretty mature, theres still the case of having to edit config files that arises on occasion but the real problem tying people to windows is applications. If your doing any task that requires workflow you generally find a very good proprietary product on Windows and an immature variant on linux. For me for example the windows killer product is cubase. There is just nothing remotely close available for linux. For others that could be photoshop or maya or whatever.
For business they have their custom application which may have been running since the 3.11 days and are still soldiering away under XP. As I said its applications that are holding back the switch.