IDEA on the chopping block ...?
Came across this blog today:
http://www.freeroller.net/page/fate/20030614#idea_on_the_chopping_block
Having read it, I can see where he's coming from. Not about IDEA, and certainly not about the people who take part in the EAP (after all, I'm one of them), but about the system of voting used to add new features.
I myself, was astounded that Aspects polled more votes than Generics. I was even more astounded to see that AspectJ support was now cluttering up the interface.
Does anyone seriously believe that by this time next year, there will be more Java developers using Aspects, than Generics?!
I've yet to see a single job advertised that asks for experience in AspectJ. As the man said, at the moment, it's an experiment, while Generics are going into the next JDK for sure.
So how did this happen?
Well the problem is much something that Apple has to struggle with just as frequently; passion. Folk here are passionate about IDEA, and with good reason. They want it to be perfect. The problem is that the folk who take part in these early access programs, like to work on the bleeding edge. They don't mind experimenting with new technology. That's why they use AspectJ.
But we are certainly not representative of the other 3 million Java developers, many of whom will not be using aspects until they are rolled into the Java; they're also not going to use a tool that interferes with compiled bytecode (I have enough trouble convincing some companies to use any code that's been auto-generated!)
I understand that JetBrains want to add what we want, but there is a much bigger picture here, and from what I've seen over the last few builds, we're not looking at it.
For years now, hardcore Apple users have been screaming for a PDA. Apple has steadfastly refused to listen to them. Why? Because the hardcore represent a small minority who will certainly buy an Apple PDA ... but the other 20million Apple users probably won't.
And this is what JetBrains has to do. Look at the bigger picture. Make an effort to find out what Java developers really want, instead of using a voting system that is always going to return a skewed view of what the Java community is really doing.
I dunno who wrote this blog though; I reckon he's reading this, and is most probably English, since use of the word 'sh*te', is something we guard very jealously... :)
请先登录再写评论。
On Tue, 17 Jun 2003 10:23:33 +0000, Jacob Bay Hansen wrote:
It amuses me somewhat, that everyone complains about other GUI builders,
and how they work, and the moment JetBrains do what they always do - bring
something fresh, and out of the ordinary - to challenge the ways you do
things, everyone still complains.
Regardless of if the final version does byte code weaving, or a standard
parent class, the concept the of their UI designer will likely stay the
same, and most likely improve.
--
...turn to the light - don't be frightened by the shadows it creates,
...turn to the light - turning away could be a terrible mistake
...dream theater - the great debate
>It amuses me somewhat, that everyone complains about other GUI builders,
>and how they work, and the moment JetBrains do what they always do - bring
>something fresh, and out of the ordinary - to challenge the ways you do
>things, everyone still complains.
I don't think, IDEA's GUI builder is better (fresher) than the other
available GUI builders, at least not in the current stage.
Tom
+So? Are we dicussing the Java Community or IDEA?
If we discuss Java, perhaps you should attack the JCP regarding Generics, and not IDEA for supporting it? While you are at it, could you slap the guy who insisted on octals in Java, and ask him why they didn't just use 0o12 (like 0xF1), and 0b0101 for that matter?+
What are you talking about ...?
+I'm using an analogy. +
Your analogy had nothing to do with what folk were talking about here. I'll assume that Octal stuff was another 'analogy'.
And YOU know better?!
I know what I like and don't like. So I express an opinion.
+Don't you fiddle with new techology? Shouldn't your IDE support that? +
Too busy doing real work, I'm afraid; but I dabble when I get the chance. As for the IDE; no, I don't use an IDE as some kind of toy. If something has a genuine benefit, then slap it in there.
"when at best, it's an experimental language" - in your humble opinion, I take it?
Nothing humble about it ...
And I think that sums up this discussion.
Good thing we have guys like you telling us the truth]]>
I'm through wasting time on you. Come back when you get wiser.
Kinda thought you were done a few messages back actually ... :)
Have a good day now ... :)
I don't think, IDEA's GUI builder is better (fresher) than the other available GUI builders, at least not in the current stage.
... but how do you make something like a GUI Builder 'fresher' ...?
My bad, I thought we were having a serious discussion, not trying to waste time.
Well, "fresher" was not my idea. I guess, Mark meant "something
new/different/more advanced/smarter".
Tom
I don't know if it's "fresh", but these guys do it kinda different: http://www.ist.co.uk/visaj/visaj2_flyer.html
There seems to be a lot of movement in the "define GUI in XML" direction, initiatives like swixml and thinlet (and many others).
apology accepted.
:)
Jacob Bay Hansen wrote:
You think it's nice???
Did you see the screenshots?
IMHO it adds just another layer of complexity. I guess that with it you could put together
HelloWorld without touching the keyboard, but I don't believe it can make me work faster.
I don't pretent to be an experienced GUI programmer, but it seems that at least they try something different. Whether they are on the right track, I don't know.
Oh, no I was having FUN reading you two bashing each other! Get back! Starting fighting again, something good may come out of it!
Good to know that. Maybe when they come out I may start using subversion :)
Perhaps we're not taking advantage of IDEA's extensibility, as much as the Eclipse group do ....
No, but we are. This is an issue that has long frustrated me. When JBuilder was first released (and likewise Eclipse) it was essentially written within their own open API. At version 1.0 it was more extensible than Idea WILL be by the time Aurora is released. It's almost as though it comes as a surprise to the IntelliJ developers that we would want to make their product better (or maybe the surprise is that we can). I love Idea, but the fact that we can't create our own live template commands, cannot create our own inspections, only very recently have been able to tap into the intention API, zero access to settings like colors/fonts, etc., had to ask repeatedly for PSI, and a litany of others strongly suggests that an "open" Idea is an afterthought at best. This is one area where I ardently wish IntelliJ would take a page out of Borland's book. They knew how to cater to people who wanted to make their product better. I think you said it best that our community will make Idea the best product possible, "if the architecture would support it."
written within their own open API.
No wonder, with several ide's behind them, they have already made many
openapi's. One more turn at doing the some thing.
Not only an afterthought, but simply something that would take time doing,
and use developer resources that where not there.
Everything else had more priority: creating new refactorings, creating
intentions, working on ease of use features, etc
Borland's book. They knew how to cater to people who wanted to make their
product better. I think you said it best that our community will make Idea
the best product possible, "if the architecture would support it."
Borland had an advantage: 10 - 15? years building ide's.
The first turbo pascal ide whas a revolution in development environments.
Sad to say, they have evolved little in the last years. Mamoth ide that does
everything, but almost no new ideas :-).
Idea is the new (r)evolution in ide's - everyone else is doing catchup
trying to copy idea features/ease of use.
I'm happy that the openapi is here (I'm really addicted to some of the
plugins), but as everything else, it had to come in it's own time.
But that's not really an excuse, is it? After all, the ideas were there for the copying.
>
But this is precisely my point! The more they work at providing an open API the less they have to do themselves. Look at the intention power pack.
Yes, and that advantage can be easily copied. And should have been.
I agree 100%. But they know how to make their IDE's open, and that architecture was there for the taking.
Yup, I agree. But Idea is playing catchup to the open architecture of many other IDEs.
I disagree. The open API is predicated on their architecture supporting it (especially with obfuscation). Therefore, the open API is rather awkward to add in later and would be best tackled and architected for at the beginning.
Dont tease :p
Ok, then maybe that's something that is worth looking at.
Thanks for the post; that was a real eye-opener. I think I understand things a little better now.
Sad to say, they have evolved little in the last years. Mamoth ide that does everything, but almost no new ideas :).
.. and is almost unaffordable ...
But this is precisely my point! The more they work at providing an open API the less they have to do themselves. Look at the intention power pack.
I think you've hit the nail on the head. Ok, AspectJ is there now. But what if someone comes up with AOP-Java? Does that mean that IDEA has to support both? What if the programmer only uses one? Shouldn't he be able to just plug in the language support he wants?
I disagree. The open API is predicated on their architecture supporting it (especially with obfuscation). Therefore, the open API is rather awkward to add in later and would be best tackled and architected for at the beginning.
Sounds daunting. ... :(
"Carlos Costa e Silva" <carlos@keysoft.pt> wrote in message
news:bcnih8$qgl$1@is.intellij.net...
essentially
>
Delphi 1.0 was released in 1995. Nothing changed much in Borland's approach
since that time. And VCL was (and is) supplied with sources.
Idea is the new (r)evolution in ide's - everyone else is doing catchup trying to copy idea features/ease of use.
That is very true. Each time there is a new Eclipse release, it has a new feature lifted straight from IDEA without so much as a name change.
But ... what happens when they do catch up? Where does IDEA go then?
Rayz wrote:
Forwards! :)
Vil.
--
__
o| . / \|o. _ _ ._ _ ._ _ |
\/ ||\/(_|| (|/||| |(/_(_)| |(/_o| |(/_ |_
/ \__
http://website.lineone.net/~vilya
For what is worth, my point is :
1. There si no doubt that idea should support latest java official specification,
2. For all that is about usability functionnalities and major module like GUI Builder, I think voting is the most democratic system and it should be the rule, as on all this subjects, it is not surprising that developpers have different opinions.
3. The problem is for all features that are aimed to support servers, framework, build systems etc.. i.e third party software or piece of code. Mostly, because the vote system is not safe (AFAIK). I hardly imagine a developper creating dozens of eap accounts just for voting "to have the satus bar in green". On the opposite, all the people behind these frameworks don't do it just for fun. It is really important to them, to be present on the market before their concurrent, and to achieve their superiority (even if it is opensource). And one way to achieve this is making a lot of noise around their framework, servers and so on.. For these reasons I think non secure votes on such "third party software" should be regarded with, at least, suspicion, and for me, should not be taken into account (as apparently a lot of people asking themselves: how can be so many vote for this ?). That's the choice of Intellij that should always prevail no matter the number of votes, until there could be some confidence about the registration (true clients only for example, with votes in rapport to their number of licences).
4. For the meticulous one : I of course exclude, from these third party software, those which are de facto standards : almost everybody use them.
Richard
Sounds daunting. ... :(
I was thinking about this some more. While architecting for an Open API from the very beginning would not be an insignificant task, I don't think the difficulty would arise from a need to over-anticipate what the Open API needs to be. It seems to me that the best approach would be for IntelliJ to design their features using their own Open API. For example, they came up with the idea of intentions. For that idea to be useful, individual intention actions must be created. It would be ideal if they originally implemented these intentions using their own Open API. Likewise with inspections, refactorings, etc. (All of which already are or are beginning to open up) This is the methodology used by JBuilder and Eclipse and why they have always been so completely extendable.
Kirk Woll wrote:
> I was thinking about this some more.
> ..
> It seems to me that the best approach would be for
> IntelliJ to design their features using their own Open API.
Well, I was thinking that JetBrains programmers work at ring 0 level;
more power, more performance, more messing-up risks. If everything was
done through the openApi, I guess they'd have to secure in completely,
to protect IDEA from our mistakes and avoid it crashing too easily
because of a trivial bug in some user-written plugin.
I may be wrong, though.
Alain
+It seems to me that the best approach would be for IntelliJ to design their features using their own Open API. For example, they came up with the idea of intentions. +
That would make a lot of sense. JetBrains pride themselves on 'eating their own dogfood', but since they have access to the core, then they may not be eating as deeply as they need to ......
I wonder if the fact that IDEA isn't as open as Eclipse, has something to do with it being obsfucated?
+This is the methodology used by JBuilder and Eclipse and why they have always been so completely extendable. +
Well, daunting as I believe it to be, it could be a ncessary course of action. Eclipse.org are planning a major rewrite for version 3.0, which aims to make the Eclipse much easier to use and also give it a nicer looking user interface.
Perhaps JetBrains will consider this for IDEA Version 5.0.
+Well, I was thinking that JetBrains programmers work at ring 0 level; more power, more performance, more messing-up risks. If everything was done through the openApi, I guess they'd have to secure in completely, to protect IDEA from our mistakes and avoid it crashing too easily +
That sounds fair enough to me; that's how they write operating systems after all.