There seems to be an assumption in the spatial Nautilus idea that real-world metaphors are a Good Thing (tm). I disagree with that -- a different medium needs sometimes a different approach.
Good examples of bad metaphors are:
Quicktime (see also the links to the RealPhone and RealCD on that page)
the desktop
and the recycle bin
To explain the latter two: the idea of the desktop was to have a central point for a document-centric environment. How many people do you know who use it that way? Most people I know use it as a pane for starting programs or just a way to have a nice background picture. I rarely see it myself since windows hide it.
The recycle bin is rather dangerous. I gave adult education classes in Windows once, and I had to learn that quite a few people empty it regularly: the full bin looks messy and they are not messy people. But that defeats the purpose of the recycle bin. (I won't go to discuss MS failure to provide this important facility where it really matters.)
The article links tries to tell me spatial Nautilus is good, because it is close to the real world. I haven't tried the new Nautilus yet, but while I actually work myself in the area of creating browsing spaces for data analysis, this particular description does not entice me at all. They can blame me for being someone who uses Windows and KDE (both true, though often Blackbox) and someone who "misuses" the browser tabbing feature (I use two windows if I have two completely different task sets -- reading Slashdot and linked sites counts as one). But that is their problem, for me the description is yet another reason not to use Gnome (the other one is that the Gnome project seems to lack pragmatism).
If they come up with a properly designed browsing space for documents (using metadata instead of tree-based hierarchies) I might be more interested.
... was this "free trade" agreement worth trading in our reasonable copyright law in exchange for selling some more sugar, wheat and wool in the US market?
Apart from the beforementioned sugar problem (which was in the mainstream media to a painful extent), I wouldn't call a copyright law "reasonable", in which taping a show on a VCR is illegal. It just makes sure that a majority of the population is a criminal.
One of the first things you should ask yourself (not Slashdot) is what your goals in open sourcing this are. That should affect not only your choice of licence (you might be limited by MySQL there) but also which SF features you want to offer.
Some ideas of what I mean, first for the licences:
is it for the greater good, and you don't mind telling people what to do (that's GPL)
you (or your company) want(s) the credit, otherwise you do not care: look at something like BSD
how much control do you want to keep over the code?
do you expect patches, do you expect to use them in a closed-world scenario? (no GPL for you)
There are many more aspects you can consider. Or you don't care. It really depends on what your goals are.
The other goal-dependend area is what you want to offer. If you just want the code out there so you can get people you know to download it: plain CVS is fine, forget the rest. Maybe upload a simple index.html onto the homepage.
If you want to reach people, you have to do more. First of all: do a decent (not fancy, but useful) website. You can get away with not updating too often (just don't put news on it if you don't plan to update), but you need a starting point for people. The SF pages and CVS are not nice, although your target audience might cope with it -- your project seems to be for rather experienced people.
How much support do you want to give? I love the support bit, helping people to use my programs gives me the kick:-) My projects thus use multiple mailing list, the webforums, a whole set of trackers (although not actively at the moment, but I will get notifications) and I try to get websites that pick interest. The latter is after all the place where I start looking at projects, and the "not there yet" often causes me to go away. Simple single page is ok, nothing is a killer unless I am really keen.
If you do not want that level of support or none at all -- turn the features off. Some contact email address is sufficient, maybe a mailing list. Noone really _needs_ web-forums and other stuff, they just help getting the casual visitors involved. Doesn't sound like your project, works for mine. Different goals, different audience --> different tools.
And a last one: SF is quite cool and they often surprise me with the quality of support you get for free; but don't expect a nice UI experience:-) The project admin bits are all over the place, sometimes you have to go into a feature section, then Admin, sometimes first Admin, then select the feature. The usual feedback is getting you back to the same page, but some more or less visible red result message somewhere on the page (mostly on top, but not always). It is not too bad, just don't have too high expectations:-)
When writing our little personal document managment tool based on Apache's Lucene, we wrote an indexer for OOo documents. Two classes: one is shared with the general XML indexer, one does the OOo specific stuff, including the extraction of metadata. In total maybe 200 SLOCs. It should handle all OOo formats if they contain text -- actually the metadata extraction should work even without.
The program also indexes Word and Excel files using Apache's POI library. I haven't looked at the size of that, but something makes me think it is a bit bigger than out little hack.
I know there is much hype around XML and in the end it is only half a syntax. But there are good applications of XML around and I think OOo is one of them.
I often used SO5.x to convert old WinWord 2.0 files into a format Office 97 could understand. SO is a quite good conversion tool for other things, too -- turning EMF into EPS, one bitmap format into the other etc.
Unfortunately it is huge and not easy to use for this, I use different tools for image conversion (mainly IrfanView) but if someone comes up with old text documents from whatever word processor -- give SO/OO a try.
You might want to take a look at this, too:
http://www.guidebeam.com/
They work on top of Google, I got this URL just some days ago and didn't find the time to check the information on their site but it might be what you are looking for.
HTH,
PeterB
Although I don't think this is a funny comment since there is some truth in it (who moderated this as funny?) you should notice that you can download a nightly built tarball of your repository -- which is probably the most important thing to have.
Losing trackers and all the other infrastructure provided would still be a nasty hit for an active project using these tools. But on the other hand: being on Sourceforge is partly so much fun since it is so huge -- I wouldn't visit or even join a large number of different sites since it would be too much overhead for me.
I would prefer to teach a programming language and not the use of an IDE. JBuilder seems much to complex to me for this use. I would like to have a kids version of VisualAge -- this is really object oriented and visual thinking.
Maybe Together might be an option -- going the UML way instead of writing classes in plain old text. Unfortunately the whiteboard edition doesn't offer the compile options but this shouldn't be that big problem with some big buttons for scripts aside.
Daddy, daddy -- why do I have to use brackets sometimes and sometimes not? Really: VB is awfully designed IMHO and shouldn't be used to teach programming. Use an OO-language not a language that has some OO-features. This dismisses C++, too. It would be like teaching word processing with MS Word -- the kids won't learn the concept of dividing content and format in Word and they won't learn goos style and the Right Way to think in VB. Peter
I don't believe an experienced assembly coder can beat or even equal an optimizing compiler in the long run. This has different reasons:
He has no time to do it
He is not able to keep track of his own optimizations -- at least not for systems of interesting size
He will make errors. To say it short: A compiler that makes errors is buggy, a programmer that makes errors is human. I can fix the compiler but not the programmer.
The compiler is build upon knowledge from many good programmers. This knowledge can increase for a long time. The programmer will reach his bounds before the compiler will. And he will eventually forgot things.
Don't misunderstand me: if a profiler tells you that there is a function consuming 10% of your CPU time and you are experienced in assembler: go for it. Compiler are not that good today. But they will be - more sooner than later. And: optimizing code is not really the creative problem. Designing code is -- if optimizing means redesigning for you, you should think on your designs earlier and maybe on a more abstract level. But maybe I'm just blinded by OOA/D/P and you're right. Time will tell. Peter
Slashdot Mirror
Good examples of bad metaphors are:
- Quicktime (see also the links to the RealPhone and RealCD on that page)
- the desktop
- and the recycle bin
To explain the latter two: the idea of the desktop was to have a central point for a document-centric environment. How many people do you know who use it that way? Most people I know use it as a pane for starting programs or just a way to have a nice background picture. I rarely see it myself since windows hide it.The recycle bin is rather dangerous. I gave adult education classes in Windows once, and I had to learn that quite a few people empty it regularly: the full bin looks messy and they are not messy people. But that defeats the purpose of the recycle bin. (I won't go to discuss MS failure to provide this important facility where it really matters.)
The article links tries to tell me spatial Nautilus is good, because it is close to the real world. I haven't tried the new Nautilus yet, but while I actually work myself in the area of creating browsing spaces for data analysis, this particular description does not entice me at all. They can blame me for being someone who uses Windows and KDE (both true, though often Blackbox) and someone who "misuses" the browser tabbing feature (I use two windows if I have two completely different task sets -- reading Slashdot and linked sites counts as one). But that is their problem, for me the description is yet another reason not to use Gnome (the other one is that the Gnome project seems to lack pragmatism).
If they come up with a properly designed browsing space for documents (using metadata instead of tree-based hierarchies) I might be more interested.
Peter
Apart from the beforementioned sugar problem (which was in the mainstream media to a painful extent), I wouldn't call a copyright law "reasonable", in which taping a show on a VCR is illegal. It just makes sure that a majority of the population is a criminal.
Some ideas of what I mean, first for the licences:
- is it for the greater good, and you don't mind telling people what to do (that's GPL)
- you (or your company) want(s) the credit, otherwise you do not care: look at something like BSD
- how much control do you want to keep over the code?
- do you expect patches, do you expect to use them in a closed-world scenario? (no GPL for you)
There are many more aspects you can consider. Or you don't care. It really depends on what your goals are.The other goal-dependend area is what you want to offer. If you just want the code out there so you can get people you know to download it: plain CVS is fine, forget the rest. Maybe upload a simple index.html onto the homepage.
If you want to reach people, you have to do more. First of all: do a decent (not fancy, but useful) website. You can get away with not updating too often (just don't put news on it if you don't plan to update), but you need a starting point for people. The SF pages and CVS are not nice, although your target audience might cope with it -- your project seems to be for rather experienced people.
How much support do you want to give? I love the support bit, helping people to use my programs gives me the kick :-) My projects thus use multiple mailing list, the webforums, a whole set of trackers (although not actively at the moment, but I will get notifications) and I try to get websites that pick interest. The latter is after all the place where I start looking at projects, and the "not there yet" often causes me to go away. Simple single page is ok, nothing is a killer unless I am really keen.
If you do not want that level of support or none at all -- turn the features off. Some contact email address is sufficient, maybe a mailing list. Noone really _needs_ web-forums and other stuff, they just help getting the casual visitors involved. Doesn't sound like your project, works for mine. Different goals, different audience --> different tools.
And a last one: SF is quite cool and they often surprise me with the quality of support you get for free; but don't expect a nice UI experience :-) The project admin bits are all over the place, sometimes you have to go into a feature section, then Admin, sometimes first Admin, then select the feature. The usual feedback is getting you back to the same page, but some more or less visible red result message somewhere on the page (mostly on top, but not always). It is not too bad, just don't have too high expectations :-)
HTH and wasn't too much of a waffle :-)
Peter
The program also indexes Word and Excel files using Apache's POI library. I haven't looked at the size of that, but something makes me think it is a bit bigger than out little hack.
I know there is much hype around XML and in the end it is only half a syntax. But there are good applications of XML around and I think OOo is one of them.
Peter
I often used SO5.x to convert old WinWord 2.0 files into a format Office 97 could understand. SO is a quite good conversion tool for other things, too -- turning EMF into EPS, one bitmap format into the other etc.
Unfortunately it is huge and not easy to use for this, I use different tools for image conversion (mainly IrfanView) but if someone comes up with old text documents from whatever word processor -- give SO/OO a try.
You might want to take a look at this, too: http://www.guidebeam.com/ They work on top of Google, I got this URL just some days ago and didn't find the time to check the information on their site but it might be what you are looking for. HTH, PeterB
Losing trackers and all the other infrastructure provided would still be a nasty hit for an active project using these tools. But on the other hand: being on Sourceforge is partly so much fun since it is so huge -- I wouldn't visit or even join a large number of different sites since it would be too much overhead for me.
--
But I wasn't able to find the RPMs for Sparc so I had to download all the ISOs :-(
Maybe Together might be an option -- going the UML way instead of writing classes in plain old text. Unfortunately the whiteboard edition doesn't offer the compile options but this shouldn't be that big problem with some big buttons for scripts aside.
Peter
Daddy, daddy -- why do I have to use brackets sometimes and sometimes not? Really: VB is awfully designed IMHO and shouldn't be used to teach programming. Use an OO-language not a language that has some OO-features. This dismisses C++, too. It would be like teaching word processing with MS Word -- the kids won't learn the concept of dividing content and format in Word and they won't learn goos style and the Right Way to think in VB. Peter
- He has no time to do it
- He is not able to keep track of his own optimizations -- at least not for systems of interesting size
- He will make errors. To say it short: A compiler that makes errors is buggy, a programmer that makes errors is human. I can fix the compiler but not the programmer.
- The compiler is build upon knowledge from many good programmers. This knowledge can increase for a long time. The programmer will reach his bounds before the compiler will. And he will eventually forgot things.
Don't misunderstand me: if a profiler tells you that there is a function consuming 10% of your CPU time and you are experienced in assembler: go for it. Compiler are not that good today. But they will be - more sooner than later. And: optimizing code is not really the creative problem. Designing code is -- if optimizing means redesigning for you, you should think on your designs earlier and maybe on a more abstract level. But maybe I'm just blinded by OOA/D/P and you're right. Time will tell. Peter