These are development tools and like it or not, developers are often asked to write visual applications, where the emphasis is on content and not beauty. I've used both perl and GD in the financial sector to plot curves, where people only wanted a visualisation of data and not a graphically designed splash of paint. I give a thumbs up to libgd.
So, I'm going to be charged for every breakpoint I set it my code? Clearly a reason to write more reliable code or step through every single line of my code; thus a reason for better reuse, shorter, efficient code and proper modularity. Would my having used break-points before not count as a prior-art?
The only reason that detecting such patterns is so hard and will in most cases need to be represented by some kind of mathematical model captured by some clever analytical engine, is because you have to examine a wide range of exemplars. My behaviour will be off every now and then and the software will need to cater for that. It's like automated trading, where you still need human input and various checks to make sure that you're not bankrupted by your software.
What happens when I come back to work with a hangover, or after my dog's gone in for surgery, or when my hands are tired from holding too many coffee cups, or my fingers unfunctional from playing too many video games or something. The day back after the holiday, when you just can't remember your password. The day after you've changed your password. In fact with every new password.
I certainly don't want to be locked out of my accounts, simply because today is a crap day and I'm feeling rather odd and out of the ordinary - as are my fingers.
It's a shame that most people seem to get blinded by the boss who talks big and employs ignorant. I know of one who holds onto the reigns of power in his own small company by controlling access to knowledge about the system which he has developed. He talks big and insults frequently. He'll argue and attempt to shame in public, even when his point is wrong. You just back down from provable truths, simply because he has his own sense of reason. Those who work with him and have any real experience are few and the bulk of his disciples are selected specifically for their lack of experience. As a result, they seem to worship him as a small deity.
The one reassuring factor is that while I have known some really brain dead heads of development in the corporate world; the further down the ladder you look the worse they get. Now based on this, things are probably not so bad. Sure, you need to be a foolish twirp to get into management, however there is still some kind of an ability requirement on the foolish twirp who lives higher up the food chain. What I find sad, is that even right up at the top, - I'm talking God or gods - well,.. do you think that S/He might also share some of those traits? I just sprained by ankle and won't provide my own opinion, S/He might over-hear me.
If you're that clueless, you probably haven't even got a unix platform to play on yet. And if you don't have a unix platform to play on yet, you should probably have a read through 'unix for dummies' first.
It's like publishing a white paper on 'uses of "the", "it" and "because" in the English language.' Kind of useless unless you already have a rudimentry understanding of the language and definately not the place to start. 'Tottaallly Poooointless (in the voice of Dr. Cox from Scrubs)" if you have already studied some English grammer.
Think about what you're proposing. API's like GWT look after all the dirties you are proposing. You're writing a rich UI which will run on multiple web browsers and cater for all their limitations. GWT takes care of this for you. You focus on your model and the essentials of your view. GWT will ensure that it will run across multiple browsers and invoke the correct remoting functions, no matter which browser it is that you're running it on. Manual porting a 'complex' interface will be an even larger pain in the rear.
That said, I know an old guy who argues that high level languages make you think less and that you should be writing everytyhing in assembler; you may also have to nibble on punch cards first. If that doesn't make you cr@p yourself, then I'd say try building a 'lots' of browsers compatible complex AJAX application totally by hand. I won't.
Ok, we have to live with software patents for the time being, but the enforcement authority is just out of line. We need lobbyists to put up the pressure on the patent office to ensure that generic classes of problems are not so easily patented.
This type of forward patenting is not of benefit to anyone, least of all Morfic, since such a train will simply reduce developer comfort when selecting technologies such as AJAX. As a developer, I'll say 'sod it' and go ahead with using something like GWT, however the odds are my employer will taboo it. This is not the way forward and it really gets my metaphorical-knickers in a twist.
I read that Microsoft held a patent on mashalling objects into XML. I usually ignore this and will go head with using an off the shelf XML writer, however I shouldn't have to feel that using such a general purpose solution will be infringing a copyright. It's not on!
It feels as though our patent office hosts a bunch of morons who will white stamp anything coming in before them. When this can have such an impact on the whole field of software development, it's simply WRONG.
Perhaps we should have a committee of forward thinkers who sit back and GPL every conceivable idea before Greedy Joe has had a chance to claim it as his own. It sickens me.
I'm now going to file a patent against generation of Javascript Basic from all high level languages which are currently mentioned in every other patent.
Very true. But those who have reached the hights of Vulcan power should be the ones who best represent the Vulcan ideal.
Te'pal continually acted as though it was that time of the month. Perhaps that's just another aspect of Vulcan physiology which I never picked up on.
So true. I'm really disturbed. Shattner, weirdo that he is, IS Kirk. Nemoy, just IS Spock. Watching Tupal on Enterprise, I just kept thinking that the franchise has messed up the whole notion of 'what' makes a true Vulcan. Roddenberry created the characters, the species, the universe and was true to it. And then along came the money moguls, and whilst they were still giving us 'good' trek, they just kept messing up the universe, which I always hoped would resemble the place where geeks go when they die. Now, I'm not 'so' sure that I still want to go there. Scotty is dead. Let him RIP. I'm sure that he's shuffling in his grave right now.
We have a whole universe to explore, why is it that they have to go back and screw with the original legend?
And yes. I would love to see a DS9 movie. If you have issues with this, it is clear that you stopped watching after the first season.
Agreed, it's BS to say you can't do arithmatic in XSLT, however, YOU SHOULDN'T.
Your stylesheet should be just for your data transformation and you don't want to be programming any complex logic into your stylesheet - unless you have to. I've seen (and written) some ugly ones and further, you don't want to be calling code from the stylesheet, which you can (depending on the xslt processor), but shouldn't - unless you have to.
Keep your stylesheets simple. They are useful, but complex logic will bloat them and it's a lot more performant to put such processing in code.
I have serious issues with XML::Simple and what it does for Perl. Sure, like everything Perlish, it makes life easier so that develoepers can quickly knock up code which performs their required task. Even bad ones. That's great. It's great if you don't know what you're doing, but it's also dangerous for the world of perl development in general.
This is my big gripe with XML::Simple. I've been using XML as protocol for dataexchange for a number of years in a number of languages and I do a great deal of Perl. I love XML::Simple when I'm parsing a very simple configuration file or a tree which is not very deep. However, it makes life so easy, that I've seen people munge in huge files with Simple, where a sax parser should have been used. I've seen people marshal out XML, when they don't understand the underlying schema - or even that such a concept exists.
A little knowledge is a dangerous thing and a (way too) powerful API is even worse. There are way too many daft Perl projects and implementations out there (in the commercial world) for this very reason. It almost makes me loath the language I love so much.
Misunderstanding. I contribute and am a oss hippie. However, it was the draw of the Darkside, which I was joking about. That said, I have gentoo 'imposed' on me and I'm not so keen on portage just yet. Further, I have no desire to become involved in maintaining the latest ebuild in a line of ebuilds, which I thought were crap to begin with - and whose predecessors I would not myself want to support.
Btw. I like whining and will forever preserve my right to whine.
What say, I just goto sf.net, download the latest stable release and build it? Why should I fix the ebuild, when I think that the whole idea of hiding possible configuration options is pants?
*I am suddenly hit by guilt <insert melancholic music>*
I should really give back to the community, shouldn't I?
Hmmm.. then again since guilt is one of the roots to the Dark side, perhaps I should simply find evil-zen in my loathing of portage?
:)
Only because it's hosted on a quantum computer, running a yet to developed version of gentoo and 14.9 linux kernel on a node on mars in the year 2132. Unfortunately, the quantum architecture is balanced with a server in Baton Rouge in the year 2003 and if you're unlucky it'll longer to serve a request than your standard windows installation; tachion latency.:(
Oh, come on. What distro, or for that matter, 'ANY OPEN SOURCE PRODUCT', doesn't come with a whole world of mailing lists, faq's and forums. Nothing special there. Move along.
It's gentoo's fault, because, sometimes you won't need generation Xi of said library. It's just imposed on you because some tosser put this dependency in the ebuild. You could probably just as easily./configure --with- to use your existing lib from source. I've seen this where pgadmin was tied to a specific version of wxwidgets and yet when building manually, it would allow you to specify which wxwidget version you were using. But noooo. Our ebuild had to force an upgrade. No it didn't, 'have to.' But it did.
I've been using Linux for over 10 years and gentoo for a couple of months now. I work in an organisation where gentoo is the defacto and I've immediately found issues with it. I've own and have worked with a number of OS's and have worked with both binary and source based packaging systems. My favourites have been bsd ports, fink, solaris packages and RPM's.
While compiling from source is will certainly give you optimisation for your given architecture, I often wonder at the situations where this should be imposed on you. I've seen people spend hours setting up gentoo users desktops and have to ask at the necessity for this. On a production environment expecting high load leves, 'yes, build from source,' however on a user desktop, I think it's waste of an organisations time and money.
Gentoo's portage system gives one the illusion of hackerish control, but having been one with hackerish control, I look at emerge and am not so satisfied. I was recently trying to build perl and wanted to fire off configure options, so I fired up emerge -vp (which shows you valid USE flags). It seems that in many places it completely curtails what you can do, depending on what the port maintainer has decided to expose in the ebuild - so I couldn't fire off half the configure options I wanted to. In my personal opinion, this leaves me wondering why I'd even want to use emerge package. Further, i was also unable to direct the build to another PREFIX directory, which is generally handy when you want to have multiple versions of components. Obviously this breaks the packaging system's world of dependencies.
Portage might be more attractive to me, if packages came as in both binary and source flavours and if there was more control over your interaction with build process. I don't like the fact that updating one emerge package seems to break your whole system and end up costing time. If I'm maintaining a desktop, I don't need this kind of hastle. I don't like emerging cpan modules, which are not consistency named. I don't need to see a large GUI application building for several hours.
SO, I ask myself, which is the best packaging system I've used. Strangely, I'm surprised, when I arrive at RPM. I've used RPM on production systems and have been surprisiingly happy with it. ebuilds are no harder than rpm spec's, however the real beauty is something which I think is essential for modern enterprise systems; transactionality. RPM v4 + allows you treat package updates as atomic transactions and in turn one can roll back from these to the previous itteration, of touched files, without having to manually manage these. SO, what happens if you break the build? You rollback.
I have a lot of friends who have used gentoo and loved it for a couple of months. I don't know anyone who hasn't shagged up his/her system and further I know a lot of these people have tired and gone back to some other distro. And we're not talking about people who didn't know what their way around a linux distro - it's typically frustration with portage. Anyway, I'm still going to give it a run for it's money and see if I end up with a different attitude.
Warning: Long and autobiographical - I got carried away
A long time ago in a galaxy far far away, I was buying a new 'computer.' This is before we only had PC's and Macs in people's homes. I'd grown up with a long line of 8-bits and this was going to be my stepping stone. So, I clearly knew what I was talking about and went and bought an ARM based Archimedes. This was a beautiful 32-bit, sexy little number which would multi-task seemlessly (for the late 80's) and as a young programmer it gave me sweet dreams at night - yes it actually did. It had - and I still miss it - an amazingly intuitive, sophisiticated and visually soothing GUI. A very powerful and novel assembly language. My only obstacles were the older programmers in my family who worked in field. "You should get a 286! This is the standard. Everyone is using it. You should use it. You'll learn something useful. It'll be forgotten and outdated soon. What is the use of learning how to program in ARM assembler anyway?';)
I bought it anyway. Not sure if I was rebelling against the status-quo, however at that time, I saw nothing cool, innovative or exciting in owning a PC. My father had one. It was a simple-but-dull bit of kit with an uninspiring OS and interesting, but hardly revolutionary applications. A fun distraction, but nothing to tun my head towards - except for Zortech C - which first exposed me to C.. I used a mac at school and could instantly see that this PC was kind of dull.
So I grew older, learned a lot with my Archi(s) and ended up doing one of my A-levels in computer science. I had a PC emulator on my archie, but it only emultated an XT and now I needed to use applications under windows 3.1. So I got a PC and installed windows. A fun novalty - for a month. RISC-OS was still leagues ahead, but I needed Windows for school. Life moved on it was kind of cool to have a machine, where you didn't need to find so niche a group of people to understand what you were talking about. It was nice to be able to get lots of games, etc, but they were quite poor. PC hardware was always fun and easy to toy with. The OS however was obviously unstable - although I'd often blame myself when life would force me to re-install. Borland turbo c++ was another toy. Still not exactly the OS of my dreams. Microsoft still hadn't sold themselves to me.
Went to UNI with my archie and eventually needed another PC. I was using sun-os predominantly during the day and my archi at home. I sold a subset of my old PC and built a new one. I was running windows 95 and it had trouble supporting old hardware, which I still had. It wasn't my favourite machine, but I had tools for some of the stuff I needed. My home kit wasn't as good, so I lived in a lab. I used my PC mostly when I went home for the holidays, and then, I'd pretty much been living in a terminal dialed into our university servers. Windows was just the glitter around it. It was unstable and occaisonally cool, but Microsoft had failed to impress me.
I didn't yet loath Microsoft. Bill Gates, as embarssed as I am to say it, was my hero. A 'geek,' as I thought then, who had made it. Well, I hadn't heard of Linus or Richard Stallman yet. And then, a friend introduced me to RedHat linux - almost ten years ago. It changed my life. I could truly work at home, contained with my 'own' UNIX environment. It was like buying an archimedes all over. Getting my box running and especially my X Server was a pain, but it was 'FUN!' - after I'd got it working. Lots of my old iso cards worked without major issue and, well, I knew that my life had changed.
Various incarnations of windows passed and I tried them. Windows 2000 struck me as being less likely to blue screen - until it blue screened. As we all do, I'd often have to sacrifice my space time to help countless very intelligent people fix their very unintuitive and temp
Was the site/.'ed or can this be the result of a DOS attack by the music industry. I believe that news of this slashdot article got through to the S0ny dirty tricks department and that they were threatened by the prospect of consumers no longer having to cyclically repurchase their albums every two years?
Or perhaps this is a result of that devious Union of Hair Gel workers, trying to avoid longer working hours.
Either way, I'm sure that the International Association of Frisbee Manufactorers is no longer laughing.
A bunch of us here, recently posed a question, which is currently intriguing me, regarding integration of eclipse with Visual Studio Team System (the ratio of.net to real-developers here is embarrassing).
As much as I aim to avoid MS, it dominates my organisation, and unfortunately the rest of the world. An expensive commercial plug-in from Teamprise was available. So this got me looking around and I found that Visual Studio Team System Foundation Server exposed a soap service for interacting with its repository. So this reminded me of an old Dr. Dobbs article which led me back to the VSTSEclipse project on sourceforge.
The project hasn't released any source yet, however looking at other mailing lists it appeared that there had been some activity in the past. I have a feeling that Teamprise actually bought up the VSTSEclipse team, given that I saw a post where a VSTSEclipse developer claimed to have started working for them. So that leaves you with the question as to who else is doing this. Which leads you to thinking that 'I'd like to write an open sourced intermediary service which communicates with VSTS Foundation Server and exposes a familiar and open interface to the rest of the world; including an Eclipse Plug-in.
Is the answer to turn American students into programming-obsessed drudges? Even if you could do that, it would just make the field less popular.
Buzzt Wrong answer. Look at the guys we look up to and those who we have worked with and know to be sh!t hot. Being programming-obsessed is critical to being a VERY GOOD programmer. And we're talking about ACM contests, there should be nothing less than the obsessed. Sure, it's healthy to have balance in life, but how many programmers are really healthy? To be an amazing computer scientist and software engineer, you should be obsessed with computer science, software engineering, mathematics or something in the ball park of our decipline.
If the standard bod obsessed with his decipline puts you off studying CS, then prehaps you don't belong in the camp, since you're not really striving or passionate about being more than a substandard computer scientist. If you respect those smelly, dirty centres of knowledge that live in the lab or are reading scientific texts for fun, then you should be encouraged to meet with them and fore-go the rich-boy party.
As I understand it, the Perl6 VM, Parrot, could potentially handle Java bytecode (after translation), thus rather than the JVM supporting P languages, we could just as easily do away with Java and start writing in Perl6. Obviously, this also implies developers having to purchase the mandatory Larry Wall shrine and Damian Conway bible. I'm sure, however, that the world will be so much better off for it. Perl 6 is due for completion in the Year of our Larry, 2666.
I usually have a commute of up to 2 hours each way. A quarter of my commute is on a railway line which runs through our remote little village [1]. The railway company does not really seem to give a wham about our little stop and the trains are regularly delayed, cancelled or lost. On top of this I am usually hopeless at getting out of bed in the morning and pull my hair out trying to get to work. I usually get to do some work on my iBook during these train journeys, but obviously no one praises me for this extra effort since there is no way to measure or recognise the dedicated service being rendered as I struggle to get in on time. Obviously, by the time I reach our office and see a room full of ties, I'm quite depressed and have lost my will to live. I often work late, but such a depressing start to the day often robs my after hours coding glory of the drive which I remember as a younger man ( two years ago ).
Now the interesting thing, which I have noticed is that when I throw the usual routine on its head and add some mutation to my search, everything works completely differently. You wake up at an insane hour of the morning and drink coffee. You then get on one of the first few trains to depart, these are invariably on time; I suppose the train drivers responsible enough to get up early are the most competent and the least likely to end up in Scotland by accident. This train is empty and free of smelly arm-pits. It is also fast and direct, requiring no further changes. Why this is not the case with the later trains, is beyond me.
The general spirit at this time of the morning, is one of champions. "I woke up before the world, therefore I am a man of power, ambition and lots of loud alarm clocks." You then stroll at leisure from station to work place with a trendy coffee in hand. The work done on the train is then casually uploaded onto workstation and you continue on a roll, glancing at those lazy sods strolling in at 9am. Your spirits are on top of the world. Come mid-afternoon, you're tired, but you've been there since the early hours. If you can't cope you can responsibly excuse yourself due to hard work and head home, stating that your work will be continued on the train. This is then valued, given that the announcement is made up front.
Somehow leaving early gives you a buzz. One should be warned, however, that insanely early starts for more than two days in a row can be hazardous to your health and lead to death by foolishly strolling in front of an old granny's very slow push bike.
[1]( George Michael lives there, and obviously never takes the train. )
i) Divide requirements into two groups: 'Core' and 'Nice to Have'
ii) Schedule a realistic release of the 'Core.' Design, develop and test your core to bits. Verify the product against the original spec. Meet the deadline and release a really robust application/device around the Core requirements.
iii) Now that we have a stable and released core, move onto the nice to have's and release a pricier, enhanced or updated version in the same fashion. Yes, some will be gutted. I was gutted when my GB Advanced looked crap compared to the SP and I was gutted when my Nokia N-gage looked stupid next to the slightly less Stupid and newer N-gage ( I like to think that these failed due to a rapidly changing market for trendy looking phones and not the quality of the product itself. ) I am both gutted and excited at the prospect of my ibook being depreciated by the newer Intel based Mac's.
IMHO, the delta between the core and 'nice to have's should be performance and quality, but 'NOT' compatibility.
What complete, total and utter nonsense. The best 'computer scientists,' whom I have worked with have been through the Eastern European system. We in the west have a tendency of learning the bare minimum to meet the criteria stated on a curriculum, which itself has been designed explicitly to allow the maximum number of students to graduate, with the minimum risk of diminishing the quality of the degree in question.
I have had close friendships with students from the Moscow Institute of Physics and Technology and I will tell you that these guys ( unless related to Mafia barons ) work their rear ends off to really understand ( over even overstand ) the Subject in question. Their curricula reflects this. Others whom I have worked with and respected have been consistent in carrying their academic knowledge out of academia and are always critical in the design of every algorithm they ever touch. Don't you wonder why there are so many Russians in Google? A multi-tier interview process, filtering the best from around the world. Annoying gits.
These guys have the hearts of hackers, but the discipline of dedicated and passionate scientists ( not even just engineers ). That said, we should remember that it is bad to stereotype.:)
Slashdot Mirror
These are development tools and like it or not, developers are often asked to write visual applications, where the emphasis is on content and not beauty. I've used both perl and GD in the financial sector to plot curves, where people only wanted a visualisation of data and not a graphically designed splash of paint. I give a thumbs up to libgd.
So, I'm going to be charged for every breakpoint I set it my code? Clearly a reason to write more reliable code or step through every single line of my code; thus a reason for better reuse, shorter, efficient code and proper modularity. Would my having used break-points before not count as a prior-art?
The only reason that detecting such patterns is so hard and will in most cases need to be represented by some kind of mathematical model captured by some clever analytical engine, is because you have to examine a wide range of exemplars. My behaviour will be off every now and then and the software will need to cater for that. It's like automated trading, where you still need human input and various checks to make sure that you're not bankrupted by your software.
What happens when I come back to work with a hangover, or after my dog's gone in for surgery, or when my hands are tired from holding too many coffee cups, or my fingers unfunctional from playing too many video games or something. The day back after the holiday, when you just can't remember your password. The day after you've changed your password. In fact with every new password.
I certainly don't want to be locked out of my accounts, simply because today is a crap day and I'm feeling rather odd and out of the ordinary - as are my fingers.
The one reassuring factor is that while I have known some really brain dead heads of development in the corporate world; the further down the ladder you look the worse they get. Now based on this, things are probably not so bad. Sure, you need to be a foolish twirp to get into management, however there is still some kind of an ability requirement on the foolish twirp who lives higher up the food chain. What I find sad, is that even right up at the top, - I'm talking God or gods - well, .. do you think that S/He might also share some of those traits? I just sprained by ankle and won't provide my own opinion, S/He might over-hear me.
If you're that clueless, you probably haven't even got a unix platform to play on yet. And if you don't have a unix platform to play on yet, you should probably have a read through 'unix for dummies' first.
It's like publishing a white paper on 'uses of "the", "it" and "because" in the English language.' Kind of useless unless you already have a rudimentry understanding of the language and definately not the place to start.
'Tottaallly Poooointless (in the voice of Dr. Cox from Scrubs)" if you have already studied some English grammer.
Think about what you're proposing. API's like GWT look after all the dirties you are proposing. You're writing a rich UI which will run on multiple web browsers and cater for all their limitations. GWT takes care of this for you. You focus on your model and the essentials of your view. GWT will ensure that it will run across multiple browsers and invoke the correct remoting functions, no matter which browser it is that you're running it on. Manual porting a 'complex' interface will be an even larger pain in the rear.
That said, I know an old guy who argues that high level languages make you think less and that you should be writing everytyhing in assembler; you may also have to nibble on punch cards first. If that doesn't make you cr@p yourself, then I'd say try building a 'lots' of browsers compatible complex AJAX application totally by hand. I won't.
Ok, we have to live with software patents for the time being, but the enforcement authority is just out of line. We need lobbyists to put up the pressure on the patent office to ensure that generic classes of problems are not so easily patented.
This type of forward patenting is not of benefit to anyone, least of all Morfic, since such a train will simply reduce developer comfort when selecting technologies such as AJAX. As a developer, I'll say 'sod it' and go ahead with using something like GWT, however the odds are my employer will taboo it. This is not the way forward and it really gets my metaphorical-knickers in a twist.
I read that Microsoft held a patent on mashalling objects into XML. I usually ignore this and will go head with using an off the shelf XML writer, however I shouldn't have to feel that using such a general purpose solution will be infringing a copyright. It's not on!
It feels as though our patent office hosts a bunch of morons who will white stamp anything coming in before them. When this can have such an impact on the whole field of software development, it's simply WRONG.
Perhaps we should have a committee of forward thinkers who sit back and GPL every conceivable idea before Greedy Joe has had a chance to claim it as his own. It sickens me.
I'm now going to file a patent against generation of Javascript Basic from all high level languages which are currently mentioned in every other patent.
Very true. But those who have reached the hights of Vulcan power should be the ones who best represent the Vulcan ideal. Te'pal continually acted as though it was that time of the month. Perhaps that's just another aspect of Vulcan physiology which I never picked up on.
We have a whole universe to explore, why is it that they have to go back and screw with the original legend?
And yes. I would love to see a DS9 movie. If you have issues with this, it is clear that you stopped watching after the first season.
Your stylesheet should be just for your data transformation and you don't want to be programming any complex logic into your stylesheet - unless you have to. I've seen (and written) some ugly ones and further, you don't want to be calling code from the stylesheet, which you can (depending on the xslt processor), but shouldn't - unless you have to.
Keep your stylesheets simple. They are useful, but complex logic will bloat them and it's a lot more performant to put such processing in code.
This is my big gripe with XML::Simple. I've been using XML as protocol for dataexchange for a number of years in a number of languages and I do a great deal of Perl. I love XML::Simple when I'm parsing a very simple configuration file or a tree which is not very deep. However, it makes life so easy, that I've seen people munge in huge files with Simple, where a sax parser should have been used. I've seen people marshal out XML, when they don't understand the underlying schema - or even that such a concept exists.
A little knowledge is a dangerous thing and a (way too) powerful API is even worse. There are way too many daft Perl projects and implementations out there (in the commercial world) for this very reason. It almost makes me loath the language I love so much.
Btw. I like whining and will forever preserve my right to whine.
*I am suddenly hit by guilt < insert melancholic music >*
I should really give back to the community, shouldn't I?
Hmmm.. then again since guilt is one of the roots to the Dark side, perhaps I should simply find evil-zen in my loathing of portage?
:) Only because it's hosted on a quantum computer, running a yet to developed version of gentoo and 14.9 linux kernel on a node on mars in the year 2132. Unfortunately, the quantum architecture is balanced with a server in Baton Rouge in the year 2003 and if you're unlucky it'll longer to serve a request than your standard windows installation; tachion latency. :(
Oh, come on. What distro, or for that matter, 'ANY OPEN SOURCE PRODUCT', doesn't come with a whole world of mailing lists, faq's and forums. Nothing special there. Move along.
It's gentoo's fault, because, sometimes you won't need generation Xi of said library. It's just imposed on you because some tosser put this dependency in the ebuild. You could probably just as easily ./configure --with- to use your existing lib from source. I've seen this where pgadmin was tied to a specific version of wxwidgets and yet when building manually, it would allow you to specify which wxwidget version you were using. But noooo. Our ebuild had to force an upgrade. No it didn't, 'have to.' But it did.
While compiling from source is will certainly give you optimisation for your given architecture, I often wonder at the situations where this should be imposed on you. I've seen people spend hours setting up gentoo users desktops and have to ask at the necessity for this. On a production environment expecting high load leves, 'yes, build from source,' however on a user desktop, I think it's waste of an organisations time and money.
Gentoo's portage system gives one the illusion of hackerish control, but having been one with hackerish control, I look at emerge and am not so satisfied. I was recently trying to build perl and wanted to fire off configure options, so I fired up emerge -vp (which shows you valid USE flags). It seems that in many places it completely curtails what you can do, depending on what the port maintainer has decided to expose in the ebuild - so I couldn't fire off half the configure options I wanted to. In my personal opinion, this leaves me wondering why I'd even want to use emerge package. Further, i was also unable to direct the build to another PREFIX directory, which is generally handy when you want to have multiple versions of components. Obviously this breaks the packaging system's world of dependencies.
Portage might be more attractive to me, if packages came as in both binary and source flavours and if there was more control over your interaction with build process. I don't like the fact that updating one emerge package seems to break your whole system and end up costing time. If I'm maintaining a desktop, I don't need this kind of hastle. I don't like emerging cpan modules, which are not consistency named. I don't need to see a large GUI application building for several hours.
SO, I ask myself, which is the best packaging system I've used. Strangely, I'm surprised, when I arrive at RPM. I've used RPM on production systems and have been surprisiingly happy with it. ebuilds are no harder than rpm spec's, however the real beauty is something which I think is essential for modern enterprise systems; transactionality. RPM v4 + allows you treat package updates as atomic transactions and in turn one can roll back from these to the previous itteration, of touched files, without having to manually manage these. SO, what happens if you break the build? You rollback.
I have a lot of friends who have used gentoo and loved it for a couple of months. I don't know anyone who hasn't shagged up his/her system and further I know a lot of these people have tired and gone back to some other distro. And we're not talking about people who didn't know what their way around a linux distro - it's typically frustration with portage. Anyway, I'm still going to give it a run for it's money and see if I end up with a different attitude.
A long time ago in a galaxy far far away, I was buying a new 'computer.' This is before we only had PC's and Macs in people's homes. I'd grown up with a long line of 8-bits and this was going to be my stepping stone. So, I clearly knew what I was talking about and went and bought an ARM based Archimedes. This was a beautiful 32-bit, sexy little number which would multi-task seemlessly (for the late 80's) and as a young programmer it gave me sweet dreams at night - yes it actually did. It had - and I still miss it - an amazingly intuitive, sophisiticated and visually soothing GUI. A very powerful and novel assembly language. My only obstacles were the older programmers in my family who worked in field. "You should get a 286! This is the standard. Everyone is using it. You should use it. You'll learn something useful. It'll be forgotten and outdated soon. What is the use of learning how to program in ARM assembler anyway?' ;)
I bought it anyway. Not sure if I was rebelling against the status-quo, however at that time, I saw nothing cool, innovative or exciting in owning a PC. My father had one. It was a simple-but-dull bit of kit with an uninspiring OS and interesting, but hardly revolutionary applications. A fun distraction, but nothing to tun my head towards - except for Zortech C - which first exposed me to C.. I used a mac at school and could instantly see that this PC was kind of dull.
So I grew older, learned a lot with my Archi(s) and ended up doing one of my A-levels in computer science. I had a PC emulator on my archie, but it only emultated an XT and now I needed to use applications under windows 3.1. So I got a PC and installed windows. A fun novalty - for a month. RISC-OS was still leagues ahead, but I needed Windows for school. Life moved on it was kind of cool to have a machine, where you didn't need to find so niche a group of people to understand what you were talking about. It was nice to be able to get lots of games, etc, but they were quite poor. PC hardware was always fun and easy to toy with. The OS however was obviously unstable - although I'd often blame myself when life would force me to re-install. Borland turbo c++ was another toy. Still not exactly the OS of my dreams. Microsoft still hadn't sold themselves to me.
Went to UNI with my archie and eventually needed another PC. I was using sun-os predominantly during the day and my archi at home. I sold a subset of my old PC and built a new one. I was running windows 95 and it had trouble supporting old hardware, which I still had. It wasn't my favourite machine, but I had tools for some of the stuff I needed. My home kit wasn't as good, so I lived in a lab. I used my PC mostly when I went home for the holidays, and then, I'd pretty much been living in a terminal dialed into our university servers. Windows was just the glitter around it. It was unstable and occaisonally cool, but Microsoft had failed to impress me.
I didn't yet loath Microsoft. Bill Gates, as embarssed as I am to say it, was my hero. A 'geek,' as I thought then, who had made it. Well, I hadn't heard of Linus or Richard Stallman yet. And then, a friend introduced me to RedHat linux - almost ten years ago. It changed my life. I could truly work at home, contained with my 'own' UNIX environment. It was like buying an archimedes all over. Getting my box running and especially my X Server was a pain, but it was 'FUN!' - after I'd got it working. Lots of my old iso cards worked without major issue and, well, I knew that my life had changed.
Various incarnations of windows passed and I tried them. Windows 2000 struck me as being less likely to blue screen - until it blue screened. As we all do, I'd often have to sacrifice my space time to help countless very intelligent people fix their very unintuitive and temp
Or perhaps this is a result of that devious Union of Hair Gel workers, trying to avoid longer working hours.
Either way, I'm sure that the International Association of Frisbee Manufactorers is no longer laughing.
It makes my hair stand. Another use for hair gel.
As much as I aim to avoid MS, it dominates my organisation, and unfortunately the rest of the world. An expensive commercial plug-in from Teamprise was available. So this got me looking around and I found that Visual Studio Team System Foundation Server exposed a soap service for interacting with its repository. So this reminded me of an old Dr. Dobbs article which led me back to the VSTSEclipse project on sourceforge.
The project hasn't released any source yet, however looking at other mailing lists it appeared that there had been some activity in the past. I have a feeling that Teamprise actually bought up the VSTSEclipse team, given that I saw a post where a VSTSEclipse developer claimed to have started working for them. So that leaves you with the question as to who else is doing this. Which leads you to thinking that 'I'd like to write an open sourced intermediary service which communicates with VSTS Foundation Server and exposes a familiar and open interface to the rest of the world; including an Eclipse Plug-in.
Does this exist yet? Anyone else interested?
Buzzt Wrong answer. Look at the guys we look up to and those who we have worked with and know to be sh!t hot. Being programming-obsessed is critical to being a VERY GOOD programmer. And we're talking about ACM contests, there should be nothing less than the obsessed. Sure, it's healthy to have balance in life, but how many programmers are really healthy? To be an amazing computer scientist and software engineer, you should be obsessed with computer science, software engineering, mathematics or something in the ball park of our decipline.
If the standard bod obsessed with his decipline puts you off studying CS, then prehaps you don't belong in the camp, since you're not really striving or passionate about being more than a substandard computer scientist. If you respect those smelly, dirty centres of knowledge that live in the lab or are reading scientific texts for fun, then you should be encouraged to meet with them and fore-go the rich-boy party.
my 0.02 EU's.
s{Che(e)rs}{Am$1n};
Now the interesting thing, which I have noticed is that when I throw the usual routine on its head and add some mutation to my search, everything works completely differently. You wake up at an insane hour of the morning and drink coffee. You then get on one of the first few trains to depart, these are invariably on time; I suppose the train drivers responsible enough to get up early are the most competent and the least likely to end up in Scotland by accident. This train is empty and free of smelly arm-pits. It is also fast and direct, requiring no further changes. Why this is not the case with the later trains, is beyond me.
The general spirit at this time of the morning, is one of champions. "I woke up before the world, therefore I am a man of power, ambition and lots of loud alarm clocks." You then stroll at leisure from station to work place with a trendy coffee in hand. The work done on the train is then casually uploaded onto workstation and you continue on a roll, glancing at those lazy sods strolling in at 9am. Your spirits are on top of the world. Come mid-afternoon, you're tired, but you've been there since the early hours. If you can't cope you can responsibly excuse yourself due to hard work and head home, stating that your work will be continued on the train. This is then valued, given that the announcement is made up front.
Somehow leaving early gives you a buzz. One should be warned, however, that insanely early starts for more than two days in a row can be hazardous to your health and lead to death by foolishly strolling in front of an old granny's very slow push bike.
[1]( George Michael lives there, and obviously never takes the train. )
i) Divide requirements into two groups: 'Core' and 'Nice to Have'
ii) Schedule a realistic release of the 'Core.' Design, develop and test your core to bits. Verify the product against the original spec. Meet the deadline and release a really robust application/device around the Core requirements.
iii) Now that we have a stable and released core, move onto the nice to have's and release a pricier, enhanced or updated version in the same fashion. Yes, some will be gutted. I was gutted when my GB Advanced looked crap compared to the SP and I was gutted when my Nokia N-gage looked stupid next to the slightly less Stupid and newer N-gage ( I like to think that these failed due to a rapidly changing market for trendy looking phones and not the quality of the product itself. ) I am both gutted and excited at the prospect of my ibook being depreciated by the newer Intel based Mac's.
IMHO, the delta between the core and 'nice to have's should be performance and quality, but 'NOT' compatibility.
I have had close friendships with students from the Moscow Institute of Physics and Technology and I will tell you that these guys ( unless related to Mafia barons ) work their rear ends off to really understand ( over even overstand ) the Subject in question. Their curricula reflects this. Others whom I have worked with and respected have been consistent in carrying their academic knowledge out of academia and are always critical in the design of every algorithm they ever touch. Don't you wonder why there are so many Russians in Google? A multi-tier interview process, filtering the best from around the world. Annoying gits.
These guys have the hearts of hackers, but the discipline of dedicated and passionate scientists ( not even just engineers ). That said, we should remember that it is bad to stereotype. :)