come to think of it... the best way to solve this is to get non-programmer types involved. Make doing the documentation not an 'also ran' task for those who couldn't hack it as developers, but as a really important part of the project - maybe we need a sourceforge solely for documentation and support forums.
UI is a bit tricky to solve in that way, but if a push to make all OSS API-driven is popular, then other people can create UIs for OSS developed software (eg. PHP front ends, windows GUIS, Java GUIs, whatever).
I don;t know about bragging rights, but certainly the sole problem is that the developers develop what they want and/or like to develop.
Documentation, feature sets, UI, these are all things that take effort and are not particularly rewarding. Hence... OSS developers don't bother too much with them.
Whereas.. for proprietary developments - you get paid to do it, no matter how crap a task it is, and if you don't, you get fired. Perhaps its because of those crappy tasks you get to do at work that many developers go home and work on OSS stuff that they like doing.
Re:UML Modelling - Communications Gap
on
UML Fever
·
· Score: 1
I'm glad to have introduced you to Crystal. I think its great.
Interesting about peer programming, I would think that every programmer would prefer to get on with it, and (if demanded) have his code inspected afterwards (as long as its by a peer, in a non-antagonistic or egotistical environment. Code reviews otherwise are a serious danger).
anyway, good luck with your projects.
Re:UML Modelling - Communications Gap
on
UML Fever
·
· Score: 1
its OK, nor am I.. err, an anti-XP zealot that is:)
I just think that it goes a little too far in its quest to be lightweight.
Peer programming is not a good idea, IMHO, as you get better 'value' from code reviews instead. (kind of like peer review afterwards instead of during - and besides, any peer programming I know of, one person codes, the other yawns, watches out the window, daydreams of Sharon in accounts, etc)
No documentation is good - as long as there is documentation produced by other means.. auto generated by code comments for example. But zero docco is a very bad thing for a project overall.
Test cases for everything before coding.. good if the design stays the same, but there is no design up front for XP, so how do you code the right stuff? Only if you've designed the test cases properly, and then write the code to fit the test - XP doesn't allow you to do *enough* design up front.
All that's my opinion. All in all, I think that if XP is 1 (on a 1..10 scale) and RUP is 10, then I'd be really happy with a methodology about the 3-4 mark.
Cockburn has some comments about XP on his crystal clear site - read his 'balancing lighness with sufficiency' article. The single best thing he says is, use the methods that work for you and your team.
Re:UML Modelling - Communications Gap
on
UML Fever
·
· Score: 1
oops. yeah, thanks for pointing that out. I do know that, its just that UML is synonymous with heavyweight methodologies like RUP in my mind - I've only ever seen it used in heavyweight methodologies, the lightweight ones use a more.. 'squiggles on scraps of paper' type design.
Re:UML Modelling - Communications Gap
on
UML Fever
·
· Score: 4, Insightful
don't forget that UML usage is driven mainly by:
a) the UML tools vendors who say its the best way to produce quality in your projects b) the people (managers usually) who believe all the stuff a) wrote.
Personally, after seeing UML used to get nowhere, I would always go for a lightweight development methodology (like XP which I dislike, or Crystal Clear)
no-one buys an item because of its LED colour, or if it has them ata ll - you buy by the spec/brand/etc. It just so happens that they all have blue leds nowadays, so telling him to buy a different product is useless advice.
Telling him how to take theproduct apart and replace the led with a different colour one... now that's the kind of answer I like.
'you know guys, you spend HOW MUCH!!!! on custom hardware? for f*cks sake, why don't you just buy some opterons?!'
Its not as if the competitive advantage of your own hardware brings much nowadays - once, when chips were slow, but today with xGhz chips being sold for peanuts, Sun might as well pack in the hardware business completely.. or devote all its money to competing in that market. I don't think it can do both hardware and software effectively.
quite true. not only that, but any software that is multi-threaded will have its performance killed not by the context switching but by software/kernel latches - things like waiting on a mutex kills performance.
So any sparc-is-better because.. drivel is just that, drivel.
well, fair enough - but there's no problem in upgrading to newer stuff, most of the time they're idle and not using up any more power than older stuff (relatively speaking - energy cost is dwarfed by the monitor, hdd, ram etc. CPU that does nothing uses next to no power). On the other hand, if you did upgrade to way powerful machines, you might not need to upgrade next time... which will mean a total cost saving.
Besides, there's a reason to use commodity hardware - cost of replacements. We had to replace a X25 card a while back.. ouch!
One thing about your reply - 'this is the most business centric IT installations'. well, yes, that's the point - IT exists to serve the business, not to serve itself.
why not - we have tax on tobacco partly to fund trhe health service (in the UK at least), and partly to stop people from smoking (to keep people healthier and therefore reduce the impact on the health service.).
There is talk of introducing a 'fat tax' on processed foods for the same reasons.
Actually, tidal friction slows the rotation of the earth and raises the orbit of the moon. Extracting tidal power will increase the friction and thus the rate at which this happens.
so the moon will get further away.
so its a secret alien plot to bring back 70s TV shows about rogue moonbases flung out to the stars.
because a terminal costs about 5 times as much a s a cheapo PC. Given that, it'll take several years for a terminal to recoup the initial expenditure based on electricity usage and administration.
Besides, if you then move to a web-based clients, the infrastructure will be in place, without spending loads on new clients.
my mate did a report on energy usage at his site, in order to cut costs blah blah blah. He said that if they just set all the PCs they had to use the power saving mode, they'd save 75,000 UKP a year.
Figure out the costs for your local company (a 17" monitor uses about 200W in use, 20W in standby) send an email to the finance director, and wait for the mainenance/admin team to rush round ensuring all PCs have their power save mode switched on.
lets put it this way.. if SCO wins anything in their legal battles, will you give $100,000 to charity? With no back-out options, no 'yeah, but it was only...' arguments, no excuses whatsoever.
Will you? Of course not, even thought you think SCO's case is complete rubbish. Why would you expect anyone else to do so then?
In the world of business, money is what matters - n-one cares what OS is used at all. The only case for Linux is the cost basis - that it is cheaper (and therefore you can squeeze higher margins, or otherwise win business that would say 'sorry, too expensive'). That's it. Linux, no-one gives a damn. Profit margin, everyone gives a damn.
except you, who have no financial stake in it at all.
don't worry - as its a Merc, it'll spend so much time in the dealer's garage being fixed, it'll have a permanent parking spot.
(for the record, I don't have one, but a very popular motoring journalist in the UK does. His comment - last christmas, he wrote 'my merc has had 2 weeks completely trouble-free, because I was in the bahamas on holiday').:)
well, you didn't expect a reasoned discussion from some people here, did you?
This is one of the other reasons business people don' care for Linux - the attitude, and excuses, of the immature geek community.
Linux, incidentally, is making inroads into some areas - web serving for example, just like Macs were the de facto DTP and image processing machines. Frankly, I can't see that changing for any reason, Linux will remain in 'niche' areas (perhaps I should say 'specialist' areas) forever, and not just for the reasons you gave (no shrinkwrapped software), but also because breaking into new niches is really difficult.
MS gives Sun some cash MS gives Sun some loose change it found lying around down the back of BG's sofa.
Sun helps MS fix.NET and user authentication problems in Window Sun accepts that a single sign on is a good thing, ushers in LibertyPassport system....
reminds me of JBA, where a cool new technology was developer internally to 'screen scrape' AS/400 screens and display them on VB forms. Great! Customers fell over themselves to get it - after all it was the best of the old tech they wanted, and cool looking desktop forms.
So, then enlisted *everybody* to perform the conversion (as the auto conversion wasn't perfect and needed hand-tweaking), and generate these VB COM dlls. 1 object per dll, for a system that had thousands of screens.
Now, if you know anything about COM, each component is a pretty heavyweight thing (its not a C++ object after all), and needs space in the registry to tell the system about itself. Registries are finite size.... well, you get the picture.
All they needed was a flyweight, but... that would have required design up-front:)
Well, they are being created right now - I saw a Horizon programme on them.
Gemesis makes yellow diamonds which are incredibly rare in nature, and makes sure that you know that they are artificial. They sell them at a 1/3rd the price of 'wild' diamonds in order not to screw the market up completely.
I guess this is how it'll be for a long, long time - they may be able to make these things for little money, but you sell them for what the market will stand, and pay.
Well, consider how many tools exist to turn UML diagrams into code, you'd think that visually-designed applications were of a much higher quality than projects hacked-up by a bunch of developers.
and I think that point came ages ago. but who gets to say when you may drop that support? yup, the Bosses - and they dont have the pain of making it work, do they.
Its a bit like if your mum patched your jeans, you wouldn't ever think how hard it was to patch them, just that they were there.
If Tridgell wants to drop support for Win95/Win3.1 from Samba, then he should do it. and live with the complaints from the people who still want the backwards compatibility.
Slashdot Mirror
come to think of it... the best way to solve this is to get non-programmer types involved. Make doing the documentation not an 'also ran' task for those who couldn't hack it as developers, but as a really important part of the project - maybe we need a sourceforge solely for documentation and support forums.
UI is a bit tricky to solve in that way, but if a push to make all OSS API-driven is popular, then other people can create UIs for OSS developed software (eg. PHP front ends, windows GUIS, Java GUIs, whatever).
I don;t know about bragging rights, but certainly the sole problem is that the developers develop what they want and/or like to develop.
Documentation, feature sets, UI, these are all things that take effort and are not particularly rewarding. Hence... OSS developers don't bother too much with them.
Whereas.. for proprietary developments - you get paid to do it, no matter how crap a task it is, and if you don't, you get fired. Perhaps its because of those crappy tasks you get to do at work that many developers go home and work on OSS stuff that they like doing.
I'm glad to have introduced you to Crystal. I think its great.
Interesting about peer programming, I would think that every programmer would prefer to get on with it, and (if demanded) have his code inspected afterwards (as long as its by a peer, in a non-antagonistic or egotistical environment. Code reviews otherwise are a serious danger).
anyway, good luck with your projects.
its OK, nor am I.. err, an anti-XP zealot that is :)
I just think that it goes a little too far in its quest to be lightweight.
Peer programming is not a good idea, IMHO, as you get better 'value' from code reviews instead. (kind of like peer review afterwards instead of during - and besides, any peer programming I know of, one person codes, the other yawns, watches out the window, daydreams of Sharon in accounts, etc)
No documentation is good - as long as there is documentation produced by other means.. auto generated by code comments for example. But zero docco is a very bad thing for a project overall.
Test cases for everything before coding.. good if the design stays the same, but there is no design up front for XP, so how do you code the right stuff? Only if you've designed the test cases properly, and then write the code to fit the test - XP doesn't allow you to do *enough* design up front.
All that's my opinion. All in all, I think that if XP is 1 (on a 1..10 scale) and RUP is 10, then I'd be really happy with a methodology about the 3-4 mark.
Cockburn has some comments about XP on his crystal clear site - read his 'balancing lighness with sufficiency' article. The single best thing he says is, use the methods that work for you and your team.
oops. yeah, thanks for pointing that out. I do know that, its just that UML is synonymous with heavyweight methodologies like RUP in my mind - I've only ever seen it used in heavyweight methodologies, the lightweight ones use a more.. 'squiggles on scraps of paper' type design.
don't forget that UML usage is driven mainly by:
a) the UML tools vendors who say its the best way to produce quality in your projects
b) the people (managers usually) who believe all the stuff a) wrote.
Personally, after seeing UML used to get nowhere, I would always go for a lightweight development methodology (like XP which I dislike, or Crystal Clear)
no-one buys an item because of its LED colour, or if it has them ata ll - you buy by the spec/brand/etc. It just so happens that they all have blue leds nowadays, so telling him to buy a different product is useless advice.
Telling him how to take theproduct apart and replace the led with a different colour one... now that's the kind of answer I like.
bollox. I think it went more like this:
'you know guys, you spend HOW MUCH!!!! on custom hardware? for f*cks sake, why don't you just buy some opterons?!'
Its not as if the competitive advantage of your own hardware brings much nowadays - once, when chips were slow, but today with xGhz chips being sold for peanuts, Sun might as well pack in the hardware business completely.. or devote all its money to competing in that market. I don't think it can do both hardware and software effectively.
quite true. not only that, but any software that is multi-threaded will have its performance killed not by the context switching but by software/kernel latches - things like waiting on a mutex kills performance.
So any sparc-is-better because.. drivel is just that, drivel.
well, fair enough - but there's no problem in upgrading to newer stuff, most of the time they're idle and not using up any more power than older stuff (relatively speaking - energy cost is dwarfed by the monitor, hdd, ram etc. CPU that does nothing uses next to no power).
On the other hand, if you did upgrade to way powerful machines, you might not need to upgrade next time... which will mean a total cost saving.
Besides, there's a reason to use commodity hardware - cost of replacements. We had to replace a X25 card a while back.. ouch!
One thing about your reply - 'this is the most business centric IT installations'. well, yes, that's the point - IT exists to serve the business, not to serve itself.
why not - we have tax on tobacco partly to fund trhe health service (in the UK at least), and partly to stop people from smoking (to keep people healthier and therefore reduce the impact on the health service.).
There is talk of introducing a 'fat tax' on processed foods for the same reasons.
no no no, read what the guy wrote:
Actually, tidal friction slows the rotation of the earth and raises the orbit of the moon. Extracting tidal power will increase the friction and thus the rate at which this happens.
so the moon will get further away.
so its a secret alien plot to bring back 70s TV shows about rogue moonbases flung out to the stars.
because a terminal costs about 5 times as much a s a cheapo PC. Given that, it'll take several years for a terminal to recoup the initial expenditure based on electricity usage and administration.
Besides, if you then move to a web-based clients, the infrastructure will be in place, without spending loads on new clients.
my mate did a report on energy usage at his site, in order to cut costs blah blah blah. He said that if they just set all the PCs they had to use the power saving mode, they'd save 75,000 UKP a year.
Figure out the costs for your local company (a 17" monitor uses about 200W in use, 20W in standby) send an email to the finance director, and wait for the mainenance/admin team to rush round ensuring all PCs have their power save mode switched on.
FYI. The original email-intercept story was reported in the Sunday Times.
lets put it this way.. if SCO wins anything in their legal battles, will you give $100,000 to charity? With no back-out options, no 'yeah, but it was only...' arguments, no excuses whatsoever.
Will you? Of course not, even thought you think SCO's case is complete rubbish. Why would you expect anyone else to do so then?
In the world of business, money is what matters - n-one cares what OS is used at all. The only case for Linux is the cost basis - that it is cheaper (and therefore you can squeeze higher margins, or otherwise win business that would say 'sorry, too expensive'). That's it. Linux, no-one gives a damn. Profit margin, everyone gives a damn.
except you, who have no financial stake in it at all.
don't worry - as its a Merc, it'll spend so much time in the dealer's garage being fixed, it'll have a permanent parking spot.
:)
(for the record, I don't have one, but a very popular motoring journalist in the UK does. His comment - last christmas, he wrote 'my merc has had 2 weeks completely trouble-free, because I was in the bahamas on holiday').
well, you didn't expect a reasoned discussion from some people here, did you?
This is one of the other reasons business people don' care for Linux - the attitude, and excuses, of the immature geek community.
Linux, incidentally, is making inroads into some areas - web serving for example, just like Macs were the de facto DTP and image processing machines. Frankly, I can't see that changing for any reason, Linux will remain in 'niche' areas (perhaps I should say 'specialist' areas) forever, and not just for the reasons you gave (no shrinkwrapped software), but also because breaking into new niches is really difficult.
I think your points 1 and 2 are not quite right..
.NET and user authentication problems in Window
MS gives Sun some cash
MS gives Sun some loose change it found lying around down the back of BG's sofa.
Sun helps MS fix
Sun accepts that a single sign on is a good thing, ushers in LibertyPassport system....
reminds me of JBA, where a cool new technology was developer internally to 'screen scrape' AS/400 screens and display them on VB forms. Great! Customers fell over themselves to get it - after all it was the best of the old tech they wanted, and cool looking desktop forms.
:)
So, then enlisted *everybody* to perform the conversion (as the auto conversion wasn't perfect and needed hand-tweaking), and generate these VB COM dlls. 1 object per dll, for a system that had thousands of screens.
Now, if you know anything about COM, each component is a pretty heavyweight thing (its not a C++ object after all), and needs space in the registry to tell the system about itself. Registries are finite size.... well, you get the picture.
All they needed was a flyweight, but... that would have required design up-front
target passersby for muggage
:)
eh? 'excuse me sir, I noticed you passing by and wondered if you'd like some hot, sweet muggage. only $1 a bowl.'
Well, they are being created right now - I saw a Horizon programme on them.
Gemesis makes yellow diamonds which are incredibly rare in nature, and makes sure that you know that they are artificial. They sell them at a 1/3rd the price of 'wild' diamonds in order not to screw the market up completely.
I guess this is how it'll be for a long, long time - they may be able to make these things for little money, but you sell them for what the market will stand, and pay.
I suppose a real diamond ring is a status symbol, as is using a $10 bill to light your cigar.
So too is throwing your money away with a virtual gift. let them who want to, do it.
Well, consider how many tools exist to turn UML diagrams into code, you'd think that visually-designed applications were of a much higher quality than projects hacked-up by a bunch of developers.
but there comes a time ....
and I think that point came ages ago. but who gets to say when you may drop that support? yup, the Bosses - and they dont have the pain of making it work, do they.
Its a bit like if your mum patched your jeans, you wouldn't ever think how hard it was to patch them, just that they were there.
If Tridgell wants to drop support for Win95/Win3.1 from Samba, then he should do it. and live with the complaints from the people who still want the backwards compatibility.