The US economy was actually in recession (i.e. negative growth) for some time during the past few years. During this period the rich got richer (i.e. their wealth grew). I'm not passing moral judgement here. I'm just pointing out that if a subset of people apparently gains money whereas the population as a whole is losing money, the proposition that the rich get richer and the poor got poorer is true.
And it makes sense that Apple is the first to ship it (they seem to pick this kind of thing up much earlier than pc manufacturers or microsoft). Bluetooth was invented for this kind of connectivity.
I think that has more to do with the requirement to have a valid key. As a consequence, people who leached the game can't play. As a consequence of that, the game servers are mostly empty save for the few weeks of hype that accompanies a game release.
A game like ut2003 stands or falls with a community. If there's a critical mass of users (legal & illegal), mods will appear, levels be written and people will just continue to play and buy the game. That never really happened with UT2003. A few people bought the game, went online and found a handfull of other people who bought the game. Then a few of them bought another game and after a few months almost nobody played ut2003 online anymore.
I understand that gamecreators want to protect their stuff. However, their actions are actually hurting their revenue because nobody buys their games after the hype is gone. What use is an online only game if the online community has moved on to the next game? Right none. Worse then hurting the revenue, they are also hurting the few people who do buy the game. These people are eager to play and after a few months their expensive game is worthless because nobody else plays it.
So here'a a suggestion. Release the game with the usual restrictions. Geeks will drool over the screenshots and buy the game no matter what. After a few months, when revenue starts to decline, remove all restrictions. By then the game will have been cracked&distrubuted anyway. Now rather than withering away, the gaming community will stay alive. You will continue to sell copies (new gamers & converted leachers) and maybe a few upgrades. This will last as long there is a community.
Quake 1 & 2 and Doom 1 & 2 continued to sell years after their release. They didn't have any restrictions. Quake 3 sold lots of copies based on the popularity of its predecessors (and the unrestricted demos that had been ciculating for months). It wouldn't have gotten that far on its own.
Yep, the guy consistently uses 'there' instead of 'their'. Actually the word 'their' does not appear in the article at all. The word 'there' is used correctly only once. Combined with the factual inaccuracies, this is a pretty lousy article.
Be-os deserves some credit for merging meta data with a file system. However, a real database goes a few steps further in terms of the ability to query, to do replication, remote data access etc.
Essentially, the Be-OS filesystem, while much richer than other filesystems, is still a filesystem. This Storage thing is a full blown SQL database in the first place.
Essentially a normal filesystem is a hierarchical database where as modern databases are relational or object databases. Relational databases have proven themselves for storing complex data over the past few years.
Some scenarios to give you a clue as to why the distinction matters: you can set up a database trigger to track changes in wordprocessor documents (i.e. automatically update some table with version info whenever you click save); you can involve external databases when doing a query on your own database (e.g. the imdb example in the Storage proposal, a tv guide); emulate a hierarchical file system by associating directory attributes with an object; emulate multiple orthogonal hierarchical filesystems; integrate security policies and encryption into the database (could also be used for DRM, I know this is a sensitive topic); make the objects themselves database records (e.g. contact information); use report generators and queries to dymamically generate complex documents (e.g. software documentation, financial overviews, etc.). Use special purpose software to browse specific types of information (e.g. a picture album, movie library or an old fashioned filebrowser).
you confuse borland with swing. I haven't used JBuilder for years and I never really liked it when I did. Probably JBuilder is slow for other reasons than just swing.
IBM may have had the vision, but the result does not exactly speak for itself. I like the eclipse GUI (well designed) but to say it is significantly better in any way than a swing gui goes to far for me. Of course there are many poorly written swing applications (one of the reasons for this is that swing is actually easy to use once you get the hang of it) but there are some good ones too.
You obviously have not been involved in EU projects where hafl of the partners is industrial (i.e. word and nothing else).
That kind of stuff imports poorly in openoffice. An word has some nice features for collaborative editing too. It totally sucks though when you have to integrate different versions of documents.
Well good for you. Actually compatibility is not that important to me. I was just observing that ooo doesn't work as advertised on my files. Crossreferences are however. That is the main reason that I'm not using it for wordprocessing yet (not the only reason though).
I'm not exactly a word fan. I'm sort of a framemaker exile at the moment (adobe seems to have more or less ceased development since 5.x and just keeps adding crap to an obsolete featureset). Word is bearable if you know where the bugs are.
Anything involving crossreferences is bound to end up in the unusable area IMHO. Openoffice currently only supports a subset of the crossreference functionality in ms office. As a consequence, editing large technical or legal documents in ooo is problematic. Sadly, the ooo developers are either not aware or indifferent to these issues (I've been all over issuezilla on this thing).
I must be a power user by the way because I have very few word documents that import correctly in ooo. IMHO ooo is perfect for the kind of stuff you could also use wordpad for (i.e. 80-90% of what business people use it for). Anything involving more complex layout stuff, embedded objects, complicated tables etc is almost certain to cause some degree of problems when importing from word. As a rule of thumb, if it needs to look good on paper don't use ooo to print a word document. If you need to do round trip editing (import, edit, export), make sure you don't lose information in the process. Both the import and export process is imperfect.
I've yet to see any swing problems that are inherent to the technology. Swing looks and feels great on my machine. I use both eclipse and swing applications and I can't notice a difference in responsiveness, look & feel, etc. I have yet to see a single swt UI for which no swing equivalent can be made. Also eclipse is not noticably faster than e.g. JEdit. It could be that your judgement is affected by the lack of hardware acceleration of swing (or 2d graphics in general) under linux. Under windows and mac os x, this is not an issue.
Admittedly it took until version 1.4.2 of the jdk for swing to catch up. I'd say swt is close to irrelevant in eclipse since it does not even include a GUI builder. Eclipse is mostly used for server side development. I think most eclipse users couldn't care less what particular toolkit is used. They just need a responsive UI and swt/eclipse happens to offer it for them.
So far the only major application to use SWT that I am aware of (no doubt there are some prototypes somewhere) is eclipse itself. I am aware of a substantial amount of mature swing apps. So to call swing a failure because swt supposedly blows swing out of the water based on a sample of one (1) application seems a bit premature. IMHO IBM wasted time and resources by developing swt. I'm sure it's a decent toolkit but I can't seem to find out what problem with swing it is trying to address or what the added value of swt is for serverside development.
There's plenty of oss players too. Bsplayer (bsplayer.org) and media player classic (http://sourceforge.net/projects/guliverkli/) come to mind. Both are excellent players and play anything you can throw at it (including DVDs if you have the right codecs installed). However, they use the ms media player codecs so they are not completely free. However, if you are on windows that is not necessarily a bad thing.
Last week a Dutch guy was caught who bribed a producer of yoghurt products. He threatened to poison products that were placed in the super market and as a demonstration placed a few poisoned products in a supermarket.
He used a US based anonymiser service to cover up his contacts with the police. He was caught because the anonymizer sevice in question happily cooperated with the legal forces, after some pressure from the dutch police and their US counterparts.
I don't approve of this guy's actions. He actually poisoned someone (who survived) with his actions. Apparently he actually tried out the poison on his goat to make sure the stuff wouldn't kill anyone. However it's a clear demonstration that anonymizers are just as anonymous as the FBI/CIA wants them to be. Anyone using the anonymizer.com services can be sure someone is watching what they do.
Re:their SE course sucks
on
MIT Everyware
·
· Score: 1
MIT is indeed a good university with excellent researchers in a number of areas. However, their reputation does not extend to the research field of software engineering where you have to look for them with a candle (well, there are some papers). Being a software engineering researcher I'm reasonably familiar with the usual channels of publication in that research field.
No doubt this course is a fine programming course. All I'm pointing out is that it has little or nothing to do with software engineering. Software engineering has little to do with technique or even the ability to program (even though that is of course a useful skill to have).
If someone is mistaken: I really like MIT's open course ware program. I'd wish more universities would do that. I was just commenting on the contents of what is cited to be its most popular course, which IMHO is labeled incorrectly as a software engineering course.
I have no comments on your remarks regarding my 'creativity'. They say more about you than me.
MIT no doubt deserves the respect you give it. And I'm well aware that they are among the elite of US universities. I just don't blindly swallow information just because of rankings (I actually like thinking out of the box). Carnegy Mellon is the place to be as far as software engineering is concerned IMHO. I myself am from the Netherlands where there are several centuries old universities with excellent reputations.
Re:Projects that work???
on
MIT Everyware
·
· Score: 1
The whole problem with software engineering in practice is that failure is very likely because many so-called software engineers don't understand their own profession. The better ones learned their trade in practice, not at a university.
The key thing to teach to students about software engineering is that programming is the easy part. Most of our students are very surprised to find out that they can't sit down and hack away from day 1. They first have to figure out what the customer wants and then come up with a realistic plan to build it. Then they find out the clock is ticking and that they actually have to build what they promised they would build.
The MIT course seems to be mainly about programming OO stuff. The requirements are fixed (very unrealistic), the customer is their teacher and the deadlines are planned for them. I'm sure the actual excercise is quite challenging but with the previous elements removed it is more about learning how to program than how to work as a software engineer with other software engineers to complete an assignment for an impatient customer.
their SE course sucks
on
MIT Everyware
·
· Score: 4, Interesting
It's a nice excercise in object oriented programming though. I've been involved in software engineering education in two universities and this is by far the least realistic course I've seen. Realism is important because otherwise students won't understand what problems await them after they finish their education. You can't teach a student to deal with the pressure of deadlines, irrational behavior from customers, customers with other priorities then you, etc. They have to experience it and be taught how to do better.
Here's how we do it (3rd year bachelor course): we group students into groups of 10, give them a contact person from a local IT company who acts as a customer and provides them with a realistic assignment (usually something that the company actually wants). Then we let them find out the hard way what software engineering is about. They have to negotiate requirements, sign a fictious contract for what they are going to deliver and then meet the terms of the contract. They have to come up with a realistic plan based on the available study points and people (i.e. 1 study points = 40 hours so 4 studypoints for the course and 10 people is quite substantial).
Meanwhile we also give them a decent introduction to software engineering (using Ian Sommerville's book, which is quite comprehensive) and make sure they understand the basics of all relevant development phases. We guide them through requirements engineering, architecture design etc.
Half way through the term after release #1, we shuffle the student groups and let them start a maintenance project on the project's first releases (i.e. you have to maintain somebody else's code with other people than during release #1).
As you can imagine this is a rather stressful period for the students but the remarkable thing is that most of them actually deliver their stuff on time, as agreed in the contract. The companies involved benefit in two ways: they get access to students who have nearly finished their education and if all goes well they get some free development time and maybe even a usable prototype. We've been doing this for a few years now and we are quite pleased with the results.
First look at this stuff: http://diveintoaccessibility.org/ http:// www.w3.org/WAI/
People already gave some advice on fonts. Here's some additional advice:
- provide an alternate stylesheet with increased font-sizes (specified in em), high contrast colors and if needed increased graphics sizes. - specify sizes of other stuff in em as well (e.g. margins and paddings). This will make sure that the content will still look good if the fonts are resized. - do not 'optimize' your site for a particular resolution.
One of the most important aspects of usability is consistency. The main problem with unix/linux is that everything is consistently inconsistent with everything else. Unix proponents call that choice, usability experts call that poor usability. Whatever it is, it seems that end users are quite capable of making their own choices and are mostly ignoring/ignorant of the linux desktop.
I recently got a Ph. D. If I had wanted to become rich instead of smart, I should have gone for a career at some.com four years ago. I didn't and I have no regrets. Getting rich is not a goal in life for me.
One thing you should realize is that if you become a Ph. D. student you are actually being trained for an academic career (plenty of jobs there if you are good enough). People in industry are generally not very interested in and rather ignorant of academia. That does not help when you are looking for a job there. It is rather frustrating to have to compete with M Sc.'s for the same job even though you are arguably more mature and educated.
Getting a job if you are smart and have the right credentials is not difficult. Getting a good job (something that actually matches your abilities) is much harder though. Typically you don't find job requirements mentioning Ph. D.'s using the regular channels (e.g. job sites). That doesn't mean there are no such jobs, it just means that you need alternative means to find them. You have to network, actively search, etc. to get to that kind of job. IMHO this is true for most good jobs.
BTW. I'm still looking for a job. Find my CV at www.xs4all.nl/~jgurp/homepage
Yep, in the Netherlands this kind of thing is long past the testing stage. My country is pretty densely populated and has quite a few highly congested highways. We have standards for sound levels which have to be met. This has created a financial incentive for road constructioners to research solutions to reducing the noise and meeting those standards.
A combination of sound deflecting shields in populated areas and better road surfaces is pretty much standard for roads nowadays here. My parents live about 1km from a very busy highway and while you can hear some noise in the background if it is really quiet (like at night) it can barely be heard.
Not bad given the market penetration of windows xp. The assumption that the crashes are distributed evenly over the userbase is of course incorrect. I had a faulty cpu a couple of months back, I saw a lot of bluescreens. I replaced the cpu and I haven't seen a single bluescreen since. Rocksolid as far as I'm concerned.
Assuming one in ten pc's has faulty hardware (bad cpu, broken motherboard, cheap memory), that translates in a lot of bluescreens. Add cheap ass hardware and flaky drivers to the mix and the numbers add up quickly.
Of course that is just hardware. Many crashes that actually get reported are usually either some windows service or explorer.exe crashing. Nasty but usually not fatal for your open applications. A reboot is the easiest way to fix but not necessarily the only way.
On the many occasions I played with linux I experienced that X is not exactly a stability wonder. Unlike an explorer.exe crash, an X crash (or even just a windowmanager crash) drags along all your open applications. Getting X to crash is not that hard. Partly due to poor hardware support, partly due to the fact that X has evolved quite significantly from its original design goals and tends to get more unstable if you start pushing your drivers to do stuff windows XP does with ease out of the box.
25 million of bsod's/day translates as: most users never see one.
Most firewall software is annoying as hell. The first step in diagnosing network problems on other machines is to tell the user to disable zonealarm or whatever piece of shit he/she is using for a firewall. You'd be surprised how often that is the source of trouble.
Firewalls are useful for servers. For workstations they generally suck and have poor usability. I haven't found one yet that is up to my standards and I've tried all major brands.
In anycase, firewalls are essential for idiots because they don't know what they are running and probably have dozens of spyware and legitimate but useless tools running. If on the other hand you do know what you are running, you also know what ports are open (namely exactly the ones that you want open). In the case of MS there are some ports you cannot close. However, you can disable most of the rpc services that use it (don't share files/printers on a external connection, kill the messenger service).
For people who know their configuration, not running a firewall is a minor, calculated risk. So what if people can ping me? So what if they can see I'm not running several familiar services? So what if they can connect to port 135? So what if Bill Gates wants send information to some vague ip addresses every few minutes?
Reuters primary goal is to provide news to other news media that generally add background information, editorials, opinions etc.
The only thing you can blame Reuters for here is that they failed to point out that Microsoft's claims are disputed by a lot of linux supporters and are not supported by independent case studies. They don't have to take sides but merely need to point out what sides there are to take in the argument.
Quoting Microsoft without highlighting the other side of the argument results in highly suggestive but factually correct reporting. However, in the context of this article it is obvious that Microsoft is trying to defend itself against what is obviously an enormous treat to their marketshare.
Thunderbird fullfills your requirements on my PC. It typically starts in about 3 to 4 seconds. This is something I don't have to do very often because I just keep it running all the time. Memory usage on windows currently is about 14 MB (It's been running all morning). I've observed that memory usage also depends on the size of the mailfolder you are currently looking at. If you open a large folder the associated index is loaded and that obviously takes some time. However, both loading and opening large folders is typically faster on my machine than similar actions in Outlook XP (which was my mail client until about 2 months ago).
I've so far not experienced a thunderbird crash even though I've been using nightlies until I installed 0.1 this morning.
XUL performs quite well on windows XP and it picks up system colors etc. The only annoying thing is that the Qute theme is a work in progress which means some of the icons are the ugly old netscape icons.
Slashdot Mirror
The US economy was actually in recession (i.e. negative growth) for some time during the past few years. During this period the rich got richer (i.e. their wealth grew). I'm not passing moral judgement here. I'm just pointing out that if a subset of people apparently gains money whereas the population as a whole is losing money, the proposition that the rich get richer and the poor got poorer is true.
Global economy certainly did not grow 10% in the past few years. You do the math.
And it makes sense that Apple is the first to ship it (they seem to pick this kind of thing up much earlier than pc manufacturers or microsoft). Bluetooth was invented for this kind of connectivity.
I think that has more to do with the requirement to have a valid key. As a consequence, people who leached the game can't play. As a consequence of that, the game servers are mostly empty save for the few weeks of hype that accompanies a game release.
A game like ut2003 stands or falls with a community. If there's a critical mass of users (legal & illegal), mods will appear, levels be written and people will just continue to play and buy the game. That never really happened with UT2003. A few people bought the game, went online and found a handfull of other people who bought the game. Then a few of them bought another game and after a few months almost nobody played ut2003 online anymore.
I understand that gamecreators want to protect their stuff. However, their actions are actually hurting their revenue because nobody buys their games after the hype is gone. What use is an online only game if the online community has moved on to the next game? Right none. Worse then hurting the revenue, they are also hurting the few people who do buy the game. These people are eager to play and after a few months their expensive game is worthless because nobody else plays it.
So here'a a suggestion. Release the game with the usual restrictions. Geeks will drool over the screenshots and buy the game no matter what. After a few months, when revenue starts to decline, remove all restrictions. By then the game will have been cracked&distrubuted anyway. Now rather than withering away, the gaming community will stay alive. You will continue to sell copies (new gamers & converted leachers) and maybe a few upgrades. This will last as long there is a community.
Quake 1 & 2 and Doom 1 & 2 continued to sell years after their release. They didn't have any restrictions. Quake 3 sold lots of copies based on the popularity of its predecessors (and the unrestricted demos that had been ciculating for months). It wouldn't have gotten that far on its own.
Yep, the guy consistently uses 'there' instead of 'their'. Actually the word 'their' does not appear in the article at all. The word 'there' is used correctly only once. Combined with the factual inaccuracies, this is a pretty lousy article.
Be-os deserves some credit for merging meta data with a file system. However, a real database goes a few steps further in terms of the ability to query, to do replication, remote data access etc.
Essentially, the Be-OS filesystem, while much richer than other filesystems, is still a filesystem. This Storage thing is a full blown SQL database in the first place.
Essentially a normal filesystem is a hierarchical database where as modern databases are relational or object databases. Relational databases have proven themselves for storing complex data over the past few years.
Some scenarios to give you a clue as to why the distinction matters: you can set up a database trigger to track changes in wordprocessor documents (i.e. automatically update some table with version info whenever you click save); you can involve external databases when doing a query on your own database (e.g. the imdb example in the Storage proposal, a tv guide); emulate a hierarchical file system by associating directory attributes with an object; emulate multiple orthogonal hierarchical filesystems; integrate security policies and encryption into the database (could also be used for DRM, I know this is a sensitive topic); make the objects themselves database records (e.g. contact information); use report generators and queries to dymamically generate complex documents (e.g. software documentation, financial overviews, etc.). Use special purpose software to browse specific types of information (e.g. a picture album, movie library or an old fashioned filebrowser).
you confuse borland with swing. I haven't used JBuilder for years and I never really liked it when I did. Probably JBuilder is slow for other reasons than just swing.
IBM may have had the vision, but the result does not exactly speak for itself. I like the eclipse GUI (well designed) but to say it is significantly better in any way than a swing gui goes to far for me. Of course there are many poorly written swing applications (one of the reasons for this is that swing is actually easy to use once you get the hang of it) but there are some good ones too.
You obviously have not been involved in EU projects where hafl of the partners is industrial (i.e. word and nothing else).
That kind of stuff imports poorly in openoffice. An word has some nice features for collaborative editing too. It totally sucks though when you have to integrate different versions of documents.
Well good for you. Actually compatibility is not that important to me. I was just observing that ooo doesn't work as advertised on my files. Crossreferences are however. That is the main reason that I'm not using it for wordprocessing yet (not the only reason though).
I'm not exactly a word fan. I'm sort of a framemaker exile at the moment (adobe seems to have more or less ceased development since 5.x and just keeps adding crap to an obsolete featureset). Word is bearable if you know where the bugs are.
Anything involving crossreferences is bound to end up in the unusable area IMHO. Openoffice currently only supports a subset of the crossreference functionality in ms office. As a consequence, editing large technical or legal documents in ooo is problematic. Sadly, the ooo developers are either not aware or indifferent to these issues (I've been all over issuezilla on this thing).
I must be a power user by the way because I have very few word documents that import correctly in ooo. IMHO ooo is perfect for the kind of stuff you could also use wordpad for (i.e. 80-90% of what business people use it for). Anything involving more complex layout stuff, embedded objects, complicated tables etc is almost certain to cause some degree of problems when importing from word. As a rule of thumb, if it needs to look good on paper don't use ooo to print a word document. If you need to do round trip editing (import, edit, export), make sure you don't lose information in the process. Both the import and export process is imperfect.
I've yet to see any swing problems that are inherent to the technology. Swing looks and feels great on my machine. I use both eclipse and swing applications and I can't notice a difference in responsiveness, look & feel, etc. I have yet to see a single swt UI for which no swing equivalent can be made. Also eclipse is not noticably faster than e.g. JEdit. It could be that your judgement is affected by the lack of hardware acceleration of swing (or 2d graphics in general) under linux. Under windows and mac os x, this is not an issue.
Admittedly it took until version 1.4.2 of the jdk for swing to catch up. I'd say swt is close to irrelevant in eclipse since it does not even include a GUI builder. Eclipse is mostly used for server side development. I think most eclipse users couldn't care less what particular toolkit is used. They just need a responsive UI and swt/eclipse happens to offer it for them.
So far the only major application to use SWT that I am aware of (no doubt there are some prototypes somewhere) is eclipse itself. I am aware of a substantial amount of mature swing apps. So to call swing a failure because swt supposedly blows swing out of the water based on a sample of one (1) application seems a bit premature. IMHO IBM wasted time and resources by developing swt. I'm sure it's a decent toolkit but I can't seem to find out what problem with swing it is trying to address or what the added value of swt is for serverside development.
There's plenty of oss players too. Bsplayer (bsplayer.org) and media player classic (http://sourceforge.net/projects/guliverkli/) come to mind. Both are excellent players and play anything you can throw at it (including DVDs if you have the right codecs installed). However, they use the ms media player codecs so they are not completely free. However, if you are on windows that is not necessarily a bad thing.
Last week a Dutch guy was caught who bribed a producer of yoghurt products. He threatened to poison products that were placed in the super market and as a demonstration placed a few poisoned products in a supermarket.
He used a US based anonymiser service to cover up his contacts with the police. He was caught because the anonymizer sevice in question happily cooperated with the legal forces, after some pressure from the dutch police and their US counterparts.
I don't approve of this guy's actions. He actually poisoned someone (who survived) with his actions. Apparently he actually tried out the poison on his goat to make sure the stuff wouldn't kill anyone. However it's a clear demonstration that anonymizers are just as anonymous as the FBI/CIA wants them to be. Anyone using the anonymizer.com services can be sure someone is watching what they do.
MIT is indeed a good university with excellent researchers in a number of areas. However, their reputation does not extend to the research field of software engineering where you have to look for them with a candle (well, there are some papers). Being a software engineering researcher I'm reasonably familiar with the usual channels of publication in that research field.
No doubt this course is a fine programming course. All I'm pointing out is that it has little or nothing to do with software engineering. Software engineering has little to do with technique or even the ability to program (even though that is of course a useful skill to have).
If someone is mistaken: I really like MIT's open course ware program. I'd wish more universities would do that. I was just commenting on the contents of what is cited to be its most popular course, which IMHO is labeled incorrectly as a software engineering course.
I have no comments on your remarks regarding my 'creativity'. They say more about you than me.
MIT no doubt deserves the respect you give it. And I'm well aware that they are among the elite of US universities. I just don't blindly swallow information just because of rankings (I actually like thinking out of the box). Carnegy Mellon is the place to be as far as software engineering is concerned IMHO. I myself am from the Netherlands where there are several centuries old universities with excellent reputations.
The whole problem with software engineering in practice is that failure is very likely because many so-called software engineers don't understand their own profession. The better ones learned their trade in practice, not at a university.
The key thing to teach to students about software engineering is that programming is the easy part. Most of our students are very surprised to find out that they can't sit down and hack away from day 1. They first have to figure out what the customer wants and then come up with a realistic plan to build it. Then they find out the clock is ticking and that they actually have to build what they promised they would build.
The MIT course seems to be mainly about programming OO stuff. The requirements are fixed (very unrealistic), the customer is their teacher and the deadlines are planned for them. I'm sure the actual excercise is quite challenging but with the previous elements removed it is more about learning how to program than how to work as a software engineer with other software engineers to complete an assignment for an impatient customer.
It's a nice excercise in object oriented programming though. I've been involved in software engineering education in two universities and this is by far the least realistic course I've seen. Realism is important because otherwise students won't understand what problems await them after they finish their education. You can't teach a student to deal with the pressure of deadlines, irrational behavior from customers, customers with other priorities then you, etc. They have to experience it and be taught how to do better.
Here's how we do it (3rd year bachelor course): we group students into groups of 10, give them a contact person from a local IT company who acts as a customer and provides them with a realistic assignment (usually something that the company actually wants). Then we let them find out the hard way what software engineering is about. They have to negotiate requirements, sign a fictious contract for what they are going to deliver and then meet the terms of the contract. They have to come up with a realistic plan based on the available study points and people (i.e. 1 study points = 40 hours so 4 studypoints for the course and 10 people is quite substantial).
Meanwhile we also give them a decent introduction to software engineering (using Ian Sommerville's book, which is quite comprehensive) and make sure they understand the basics of all relevant development phases. We guide them through requirements engineering, architecture design etc.
Half way through the term after release #1, we shuffle the student groups and let them start a maintenance project on the project's first releases (i.e. you have to maintain somebody else's code with other people than during release #1).
As you can imagine this is a rather stressful period for the students but the remarkable thing is that most of them actually deliver their stuff on time, as agreed in the contract. The companies involved benefit in two ways: they get access to students who have nearly finished their education and if all goes well they get some free development time and maybe even a usable prototype. We've been doing this for a few years now and we are quite pleased with the results.
First look at this stuff:/ www.w3.org/WAI/
http://diveintoaccessibility.org/
http:/
People already gave some advice on fonts. Here's some additional advice:
- provide an alternate stylesheet with increased font-sizes (specified in em), high contrast colors and if needed increased graphics sizes.
- specify sizes of other stuff in em as well (e.g. margins and paddings). This will make sure that the content will still look good if the fonts are resized.
- do not 'optimize' your site for a particular resolution.
One of the most important aspects of usability is consistency. The main problem with unix/linux is that everything is consistently inconsistent with everything else. Unix proponents call that choice, usability experts call that poor usability. Whatever it is, it seems that end users are quite capable of making their own choices and are mostly ignoring/ignorant of the linux desktop.
I recently got a Ph. D. If I had wanted to become rich instead of smart, I should have gone for a career at some .com four years ago. I didn't and I have no regrets. Getting rich is not a goal in life for me.
One thing you should realize is that if you become a Ph. D. student you are actually being trained for an academic career (plenty of jobs there if you are good enough). People in industry are generally not very interested in and rather ignorant of academia. That does not help when you are looking for a job there. It is rather frustrating to have to compete with M Sc.'s for the same job even though you are arguably more mature and educated.
Getting a job if you are smart and have the right credentials is not difficult. Getting a good job (something that actually matches your abilities) is much harder though. Typically you don't find job requirements mentioning Ph. D.'s using the regular channels (e.g. job sites). That doesn't mean there are no such jobs, it just means that you need alternative means to find them. You have to network, actively search, etc. to get to that kind of job. IMHO this is true for most good jobs.
BTW. I'm still looking for a job. Find my CV at www.xs4all.nl/~jgurp/homepage
Yep, in the Netherlands this kind of thing is long past the testing stage. My country is pretty densely populated and has quite a few highly congested highways. We have standards for sound levels which have to be met. This has created a financial incentive for road constructioners to research solutions to reducing the noise and meeting those standards.
A combination of sound deflecting shields in populated areas and better road surfaces is pretty much standard for roads nowadays here. My parents live about 1km from a very busy highway and while you can hear some noise in the background if it is really quiet (like at night) it can barely be heard.
Not bad given the market penetration of windows xp. The assumption that the crashes are distributed evenly over the userbase is of course incorrect. I had a faulty cpu a couple of months back, I saw a lot of bluescreens. I replaced the cpu and I haven't seen a single bluescreen since. Rocksolid as far as I'm concerned.
Assuming one in ten pc's has faulty hardware (bad cpu, broken motherboard, cheap memory), that translates in a lot of bluescreens. Add cheap ass hardware and flaky drivers to the mix and the numbers add up quickly.
Of course that is just hardware. Many crashes that actually get reported are usually either some windows service or explorer.exe crashing. Nasty but usually not fatal for your open applications. A reboot is the easiest way to fix but not necessarily the only way.
On the many occasions I played with linux I experienced that X is not exactly a stability wonder. Unlike an explorer.exe crash, an X crash (or even just a windowmanager crash) drags along all your open applications. Getting X to crash is not that hard. Partly due to poor hardware support, partly due to the fact that X has evolved quite significantly from its original design goals and tends to get more unstable if you start pushing your drivers to do stuff windows XP does with ease out of the box.
25 million of bsod's/day translates as: most users never see one.
Most firewall software is annoying as hell. The first step in diagnosing network problems on other machines is to tell the user to disable zonealarm or whatever piece of shit he/she is using for a firewall. You'd be surprised how often that is the source of trouble.
Firewalls are useful for servers. For workstations they generally suck and have poor usability. I haven't found one yet that is up to my standards and I've tried all major brands.
In anycase, firewalls are essential for idiots because they don't know what they are running and probably have dozens of spyware and legitimate but useless tools running. If on the other hand you do know what you are running, you also know what ports are open (namely exactly the ones that you want open). In the case of MS there are some ports you cannot close. However, you can disable most of the rpc services that use it (don't share files/printers on a external connection, kill the messenger service).
For people who know their configuration, not running a firewall is a minor, calculated risk. So what if people can ping me? So what if they can see I'm not running several familiar services? So what if they can connect to port 135? So what if Bill Gates wants send information to some vague ip addresses every few minutes?
Reuters primary goal is to provide news to other news media that generally add background information, editorials, opinions etc.
The only thing you can blame Reuters for here is that they failed to point out that Microsoft's claims are disputed by a lot of linux supporters and are not supported by independent case studies. They don't have to take sides but merely need to point out what sides there are to take in the argument.
Quoting Microsoft without highlighting the other side of the argument results in highly suggestive but factually correct reporting. However, in the context of this article it is obvious that Microsoft is trying to defend itself against what is obviously an enormous treat to their marketshare.
Thunderbird fullfills your requirements on my PC. It typically starts in about 3 to 4 seconds. This is something I don't have to do very often because I just keep it running all the time. Memory usage on windows currently is about 14 MB (It's been running all morning). I've observed that memory usage also depends on the size of the mailfolder you are currently looking at. If you open a large folder the associated index is loaded and that obviously takes some time. However, both loading and opening large folders is typically faster on my machine than similar actions in Outlook XP (which was my mail client until about 2 months ago).
I've so far not experienced a thunderbird crash even though I've been using nightlies until I installed 0.1 this morning.
XUL performs quite well on windows XP and it picks up system colors etc. The only annoying thing is that the Qute theme is a work in progress which means some of the icons are the ugly old netscape icons.
they probably left out some #define's that make it valid C :-)