What is the user community's preference for where JetBrains should focus?

I've been an IDEA user and customer for several versions now. Since the company I work for doesn't pay for Java IDEs, I've been buying it personally.

Version 6.0 is the first time I didn't adopt it immediately (I am now evaluating it). The reason I held off was because of all the seeming dissent among users.

I am curious to know whether the user community at large is demanding all the new features... Personally, I would actually pay for a release where JetBrains only reduces the memory footprint, and makes the environment snappier. Really modular or whatever.

Yes, we're talking about a relatively small program but for us poor folk who lug around laptops that run a database and a mammoth appserver, every meg counts. Am I an outlier, or is this a common sentiment?

Regardless, thank you for the IDE; thank you for being really screen space concious.

Comment actions Permalink

Firstly, credits to you for forking out your hard earned to pay for the IDE. I hope you're employer realises how lucky they are. Secondly, I'd be very reluctant to work in such an environment.

Re IntelliJ. We've recently upgraded, and like you I've been an IntelliJ user for years (since 2001) and also held off on the upgrade. We have now upgraded, and I'm still finding new features that are very nice, just little things, but things that make my day better.

Finally - RAM is cheap, you should be able to stick 2gig in a laptop. I'd do that and be done with it. I run oracle, weblogic, intellij and outlook inside 2gig. So to answer your query - I'm happy with effort being put into features rather than streamlining the app.


Comment actions Permalink

Most new features being put into the releases I don't use. For me it would be nice if JetBrains put some work into the performance of IntelliJ. The startup time is sort of ridiculous...especially the first start of IntelliJ after restarting my computer...that takes forever!

I do have to say though that last week I upgraded my laptop to 2 Gigs of Ram (dual core 2ghz processor) and most of the performance pain has gone away...I can now run IntelliJ and a local Weblogic instance at the same time and it isn't painful...it is actually pretty snappy.

Now if the "Persistence Units" supported Hibernate that IS a feature I would use (as far as I can tell it is only for Entity beans right now). Currently for hibernate I use a combination of middlegen and hand editing the mapping files.

Comment actions Permalink

MP> Most new features being put into the releases I don't use. For me
MP> it would be nice if JetBrains put some work into the performance of
MP> Now IS a feature I would use

The usual 20%/80% thing. Why is the company wasting it's time working on
those stupid features I don't have a use for, instead of focusing on performance
and this special feature that I really need just now (TM).

For what' it's worth, Jetbrains: stop working on useless features [j2ee,
j2me, html, javascript, hibernate, structs, ruby, all VCS other than subversion,
etc] and focus on performance and rich client features, which is what I use.
Great waste of resources most of the improvements since idea 3.0 ;)

Comment actions Permalink

I would say to anyone: "Upgrade now!". 6 has some extremely nice features and is quite solid now. I don't have any performance problems with an 8 month old laptop (dual core with 2 GB ram). There really isn't any other product that can compare with Idea in terms of productivity.

Comment actions Permalink

I would say without hesitation, performance needs to be addressed. The only time I post on this board is when I'm waiting for IntelliJ to finish some operation. I usually have plenty of time to finish my post!

All of the feedback mechanisms (progress bars, etc) to indicate what's IntelliJ is doing or how far it is going in an operation seem to be broken. Things progress to 99% in 2 seconds and then stay there for 5 minutes. This is extremely frustrating from a users perspective. Intellij has an annoying habit of being slowest when an Eclipse user is around, to which they reply "Why are you using this thing?". It's getting harder and harder to come up with a good reply ... :(

Comment actions Permalink

For myself, I want what I've always gotten from IDEA: 1) butter smooth usability 2) deep knowlege of code structure and 3) an understanding of what developer workflow is actually like, rather than an idealized-to-the-point-of-nonsense view of what it should be like (I'm looking at you, Rational!).

I consider 1) and 2) basically feature-complete as of IDEA 5.1 (or 6.0 for web developers), although there is always room for polishing and understanding more languages. 3) still holds a lot of room for feature innovation, though, particularly with TeamCity to leverage. Clear management of multi-branch projects, issue tracking and code review integrations, project level reporting, richer testing and coverage functionality, tracability, integration of system monitoring tooling, the cool "shelve changes" stuff in 7.0, etc. Borland was trying to brand all this as "Application Lifecycle Management", but they don't really have any technology to back the promise of the idea. 6.0 showed quite a bit of improvement int this area (code coverage, TeamCity, VCS improvements), and that's basically why I recommend the upgrade (unless you're doing web development, in which case there's a lot of reasons to upgrade).

(Oh, yeah. Buy more memory. Having less than 2 Gigs on a developer machine nowadays is just asking for heartache. That would be true even if you were using vi as your editor.)

--Dave Griffith

Comment actions Permalink

Another user since 1.0 here.

I have bought IDEA for myself a couple of times. I personally am probably responsible for converting over a 100 developers to IDEA. When I started at my current employer (Agitar) I was able to convert ALL of the developers to IDEA. One by one they have switched to Eclipse. I am the last hold out.

I evaluated 6.0 and, for the first time ever, I have not upgraded immediately . I know my boss will say - everyone else switched to Eclipse. Why don't you? For the first time ever, I can't say "because IDEA is so much better".

Each release adds a bunch more 'newbie' features that I don't find particularly useful - GUI Builders, wizards for EJB and the like. What attracted me to IDEA in the first place was that it focussed on the core things that a professional developer needs. Over the years, I have felt that the target audience for IDEA is drifting further and further to the left along the bell curve and the professional developer is being neglected more and more.

I appreciate all the support for the various web-related technologies (CSS, JS, HTML, XML etc) but I have never found a GUI Builder that works for anything more than a trivial app - including the builder in 6.0. The wizards for EJB give up way to easily - I'll edit the xml thanks.

+1 for a focus on performance. Startup time is outrageous. I get coffee, read, email, surf the web and much much more while IDEA opens my project.

In 6.0, the code coverage is nice and I like the JUnit4.0 and SVN support but they don't quite justify the upgrade price for me. If you ever have one of those personal edition ($100) promotions, I might buy it for myself but I doubt I'll ask my employer to shell out the big bucks. I'll just struggle along with 5.0. It's all very sad.


You send us code. We send back tests. For free.

Comment actions Permalink

Hi Kevin,

I don't know why so many people are whining about startup time. We have a medium sized project (corba enterprise app with swing client, > 5000 classes, > 6000 files, no JEE appserver, no web stuff), which is loaded in about 1.5 minutes. Not perfect, but absolutely usable. I load IntelliJ once a day, so this really does not count. So unfortunately, I even cannot finish e-mail reading, coffee drinking, and web surfing until IntelliJ finishes loading. So, Jetbrains, please slow down project loading so that I find time to do all this stuff ;) .

Seriously, without describing what kind of project you are working on, it does not make sense to complain about startup performance. I would prefer, if JetBrains could maintain startup performance while adding features, instead of putting a lot of resources into this issue on the expense of adding features in the field of Hibernate and Spring support or adding productivity enhancements to VCS integration.

Although we don't use IntelliJ's gui builder (we use JFormDesigner plugin), I value the availability of a GUI Builder very high. We use GUI Builders not only for the production dialogs, but also for prototyping and specifications. In those cases a visual GUI Builder is much quicker, produces are more realistic view as compared to Visio scetches, and can even be used by a person not familiar with Java (JFormDesigner is also a standalone product).

Ultimately, until now each IntelliJ release had enough features (better svn integration and web development features, javascript etc.) to justify the upgrade. In its basic functionality IntelliJ is feature-complete, so the productivity gains may become smaller with each release. However, if we would not buy the upgrades, JetBrains would possibly not survive and IntelliJ might disapper so that we would have to switch to Eclipse, which would be the real pain. In my opinion, the value for money is still there, so no reason to complain.

Kind regards

Thomas Gülden
Munich, Germany

Comment actions Permalink

The startup time really depends very little on the nature of files contained in the project, but rather on the number of files: the reason is that IntelliJ scans through all the files trying to find modified/deleted/added ones. The activity is mostly disk-reading, and keeping disk defragmented becomes crucial.


Comment actions Permalink

Yes, the basic functionality is indeed very much complete. For Java. To make development more pleasureable, one needs to either advance Java language expressiveness or face the fact that there could be more than one language targeting JVM used (think JRuby, Groovy, Scala). Only when the community is ready for this step (and the language designers are ready to stabilize their languages), will the IDE makers be able to make advances.

Comment actions Permalink

Agreed. I'm very much looking forward to using (and writing!) closure conversion inspections/intentions/refactorings if and when closures actually get added to the language. Until then, supporting more of the little "languages" that have accreted around Java (Spring, Hibernate, Tapestry, Maven, etc. ) is a very worthwhile step.

Alternatively, you could finish/open the Scala plugin, and see if the industry is willing to make that jump...


Comment actions Permalink

I have been using the latest Intellij and honestly there are no performance problems. This on a machine with 2.8 ghz and 1 gig of RAM. I have 14 modules loaded in the project and each module contains no les sthan around 300 files I am guessing. When Intellij starts up it does take a bit of time, but that is because the hard drive is quite slow on this machine. Probably 5400 RPM. Intellij is running quite zippy. I have Intellij setup to run on Java 6 and it runs fine. The max memory footprint I have seen this IDE grow to is about 350 MB.

Intellij is very disk intensive IMO so I defrag my work machine once a week to keep the performance up.

I see what a lot of people are saying as to focus more on professional java developers rather than noobs and I agree with that a lot. Wizards are fine as long as they are very powerful sort of like the ones JDeveloper have. So far I am quite happy with Intellij. However, having used the latest JDeveloper, I must say Oracle has done an amazing job keeping the performance so high in the IDE even with all the features included! I have to say Intellij in comparison has a long way to catch up to that.

Comment actions Permalink

Dave, we are on the way to open the first version of Scala plugin. This is not going to be anything comparable to Java support in IDEA, but we'd like to get the feedback (and conributions:) already at this early stage.

Comment actions Permalink

Performance then usability are the biggest concerns. If, for example, you use a typical corporate samba mount and a decently sized project, IDEA was not practical to use until maybe 6.0.3. IDEA 6 continues to exhibit performance under certain circumstances that make it unpleasant to use. If you have a certain setup, unfortunately, IDEA performance has been poisoning people's opinion of the product. Here's an example issue that basically makes IDEA unusable if you encounter it:

From the perspective of JetBrain's management & PR, I'm not sure why such issues don't get immediate action and timely resolution given the negative impact they have on IDEA's performance and IDEA's reputation.

Comment actions Permalink

I am still sitting on 5.1 and will not be going to 6.0.

I am now playing with the new EAP and will waiting and see on it. Though, ironically, it is usually the available plugins that are usually the final piece that allows me to justify the upgrade.

However, this time around, unless 7.0 and the plugins blow my socks off, I will also be re-evaluating Eclipse, the new Netbeans, and WSAD(which we are supposed to be using here, I have a license, but just have not gotten about to installing it). That again, I may just sit at 5.1, since it is getting the job done for me, though there appear to be a few bugs have been fixed in 6.0 or 7.0; but I hate paying for bug fixes.

Overall, I still look for the small things, that impact me more often or that remove hassles, like being able to create live templates that start with a /(without the need to shut down IDEA, edit the XML, and then restart), a date variable for live templates, or more flexibility in the Extract Method refactoring.

Don't get me wrong, I still like IDEA and it has helped me in many ways, but I won't blindly follow the upgrade trail.

Comment actions Permalink

Wheee! Is there a separate forum for discussing this? Or a dedicated
thread? Any idea about a date for a 0.1 version?

Comment actions Permalink

What about the MPS project?

Comment actions Permalink

As I see it, in buying an upgrade, we are basically paying for continued IntelliJ support. MyEclipse make this charge very clear, either $29 or $49 per year for the ability to use the product. If you don't pay, it stops working..

Before upgrading to 6.0, I tried Eclipse 3.2 with both WTP and MyEclipse and stuck with it for several months, there were some good points, CVS / SVN synchronization support is excellent - but I still prefer IDEA for day to day use.

Comment actions Permalink

And for others the set of features they consider useful is quite different.

For me the JEE stuff is extremely important (especially JBoss integration).
Subversion I couldn't care less about though as we don't use it, we use CVS.
Struts, Ruby, and JME are not important for me either, but other projects here do use Struts and I'm sure there are people out there using JME and Ruby (though why you'd infect a Java IDE with Ruby (or indeed why people would use Ruby, I looked at it and discarded it almost instantly) is beyond me).

Performance is quite decent, especially compared to Eclipse and Netbeans, so not much to complain about there.
Only thing bothering me is the frequent rescanning of build directories after running an external ANT task, which can take minutes.
Would that could be turned off, or selectively applied (I don't need it to constantly scan doc and binary directories for changes).

Comment actions Permalink

Some of the features of the latest release are great but most are useless to me.
I can not event point to one feature that made THE difference for me.. (may be Changes tab but I used Tmate while it was alive (WTF happen to it!!!?!?!?!?!?! so the functionality is not new)

I have the moderate size project with a mix of java, jni and xml. There is still no support for C++ (i have to use eclipse for that) and the speed/stability is awful !!!!

I am seriously considering to switch back to eclipse which is closing the feature set gap very quickly and has some build in functionality that MUST be in IDEA but was never implemented (although ppl requested it for a few years now.)

the simplest example is the patch process. My team works with patches. We do not commit changes to the production branch of VCS. Eclipse, NetBeans and JBuilder have build in support for creating/applying patches and IDEA does not.... I am wasting a lot of time in cygwin....

Comment actions Permalink

Hello Yaroslav,

the simplest example is the patch process. My team works with patches.
We do not commit changes to the production branch of VCS. Eclipse,
NetBeans and JBuilder have build in support for creating/applying
patches and IDEA does not.... I am wasting a lot of time in cygwin....

Creating and applying patches is already implemented in Selena. Feel free
to try out the EAP build:

Dmitry Jemerov
Software Developer
JetBrains, Inc.
"Develop with Pleasure!"


Please sign in to leave a comment.