You know between Java, Hudson and moves like this it feels like we're ripe for an Onion-style article about Oracle execs running out of ideas for pissing off the OSS community and so doing a road-trip around the country to individually insult every OSS developer and user in the most graphic and personal way they can think of.
Oracle has a profitable business model. It works for them. They have some open source projects that they no longer care to support. In the open source model, someone else is free to carry the torch. It should only require a few people to adopt the Ruby plugin. There are already lots of Netbeans plugins so the learning curve:
A fair point for Rails-on-Netbeans. Not quite so much for Hudson, where Oracle wants not just to stay involved, but to stay in control --- and the amount of control they want is what's pissing off everyone else (including the primary author of the software, Kohsuke Kawaguchi).
So it's actually better for the Ruby/Rails plugin to be dropped by Oracle. Then the community doesn't have to fork to save it from Oracle's controlling hands.
Anyone who cares about open source doesn't buys oracle. So now it doesn't matter. 5 years from now, when "open source believer" rise to a management position, and all the infrastructure and people is oracle, the is really very little decision power.
The OSS community for Hudson (actually, CloudBees) is forking Hudson against Oracle's will.
RoR support is just dropped due to various constraints and the community is actually encouraged to intervene:
>However, we strongly encourage our community of NetBeans Ruby users and developers to volunteer to take on development of Ruby on Rails support for the NetBeans IDE.
I'm curious if people will pick up RoR support or just continue complaining about Oracle because they don't make things for free.
I've used it for (J)Ruby GUI development. The built-in WYSYWYG editor is pretty sweet, so Netbeans works very well as as an all-in-one dev tool for building JRuby + Swing + Monkeybars apps.
I've never used it for any Rails development, and found the coupling of the Ruby packages with Rails gems, and the presumption in the add-ons that any Ruby project was a Rails project, a bit dopey. There seems to be no way to install the Ruby stuff without the Rail stuff, though.
I much prefer using vim for development, and while there's a vim plugin for Netbeans I typically did serious editing in an external vim instance while using Netbeans for GUI stuff, or for some refactoring, which it does quite well.
For people who like IDEs, Netbeans is a really good Ruby/JRuby editor.
I've also used for several sinatra apps. Their refactoring, autocomplete, debugging and unit test/rpec is superb. Not to mention it supports Mercurial and is free, what Intellij's Ruby IDE doesn't (the core is open and free, and supports mercurial, but their ruby thing is not a plugin, and it is not compatible with the good stuff). Yeah, its a loss, but they've been neglecting it for a while, so it is sad.
Just a quick "me too" here. It's a great combination for knocking together a quick GUI around whatever random hack the business needs today. I don't know of a Swing editor that's as good.
"Second: Although our Ruby support has historically been well received, based on existing low usage trends we are unable to justify the continued allocation of resources to support the feature."
I would say, not a whole lot according to their trend analysis.
I don't agree with some of the things Oracle has done as of late, but that seems quite reasonable to me. If usage levels are quite low, then it is a waste of time to keep devoting resources to it.
Anecdotally, I don't know of a single RoR developer who uses Netbeans. Vim, Textmate and Emacs seem to be the editors of choice.
I've been using NetBeans on Linux as my main Rails IDE for the last few years. I kinda like netbeans, but I've been using it since about v4 so I'm somewhat accustomed to it.
Features I make use of: project navigation, file navigation, syntax highlighting, jump to a declaration, re-indent and very occasional code-complete (mostly for css)
Features I don't use: everything else. I always have a terminal window open with about 5/6 tabs from which i do all my generations, migrations, console, server etc. I use RVM and NetBeans doesn't seem to play nicely with that
I do JRuby/Rails development. Typically I'm in emacs (Aquamacs, actually), but I've used NetBeans occasionally when I wanted to use the debugger... beats the hell out of tracing w/ Rails.logger.
I don't understand why this move should piss off anybody in the OSS community. Oracle is a private company and as such has only obligations to its shareholders. They may or may not support and/or promote OSS projects as they see fit, same as every one of us. It's their free choice.
Look at what kind of headway corporations like google have made in the mind of consumers when they go above and beyond the shareholder mantra. Oracle might make short term gains on their terms but they have started to poison a whole new generation of developers.
Webpages are not a complement to Oracle's business model, thus what makes sense for Oracle and Google to open source are vastly different. If you look at what Google open sources, it's all compliments to their business model. Google doesn't open source technology that it makes money from.
As to going above and beyond in regards to open source, you'd have to look at where NetBeans came from Sun. Look where all that open source got them.
Oracle is not a person to make such a decision; it is its shareholders who should decide how they want to spend their money. I wouldn't want to have a tax (as an Oracle investor or a private citizen) to support OSS. I may want to contribute my money and/or code as a donation, but that will be my own choice.
The reason is that Oracle inherited the stewardship of so many important projects when they bought Sun. After the purchase they have not said 'OK guys, new management; we don't do the open source thing; so long.' Rather they put the knife into various projects in different ways, seemingly at random. This behaviour taken in sum has an unsettling effect on the outlook of any project that Oracle is still involved with. How long will Netbeans itself survive? It's anybody's guess.
It's not just one case of support being withdrawn. It's Russian roulette.
That they are free to make this choice doesn't mean the community needs to roll over in accepting it. A response to this news that doesn't indicate displeasure would lack integrity.
I'll try to read that over the weekend, but I'll just note that it's not difficult, especially for someone of Friedman's stature and intelligence, to construct an internally consistent exposition of a particular worldview. That is to say, I hope he accounts for falsifiability. ;)
> To maintain that objective and capitalize on the JDK 7 release themes [...] it is necessary that our engineering resources are committed to [...] NetBeans IDE 7.0.
(from http://netbeans.org/community/news/show/1507.html )
and
>However, we strongly encourage our community of NetBeans Ruby users and developers to volunteer to take on development of Ruby on Rails support for the NetBeans IDE.
(from http://wiki.netbeans.org/RubySupport )
So it's just a matter of not having enough resources to support everything.
Ruby and RoR is done basically entirely in-house by NetBeans/Oracle so it seems odd that people are enraged that something they provided for free isn't supported any longer due to normal restrictions.
I'm curious if:
1. People outside Oracle will take over RoR development (the community seems too small) OR
2. People will pledge money to keep the plugin maintained (I somehow doubt this but I would like to see it happen).
I think you're right. The intersection of interested Ruby developers + with Java experience + knowledge of the Netbeans platform + and motivation/time to maintain the plugin seems vanishingly small.
The only thing is that Ruby is still growing, and so represents possible future revenue. But Oracle aims for short-term profitability now... which you have to admit, has worked out pretty well for them.
Taking a wider view, I love the work to make things better (such as at Xerox PARC), but it hasn't made money for those companies. Which is partly why Sun no longer has any say in Java, or in general research, or indeed in anything at all.
Thanks for the tip - will be checking this out tonight. Double thanks for the MacVim tip - I'm sure I would have been tripped up and off to Google without it.
I have heard positive things said about RubyMine from JetBrains. However, it is not free and some complain of slow project parse times at startup. Disclaimer: I do have an off and on professional relationship with JetBrains.
RubyMine is amazing for large Rails projects. I can jump between methods in different modules, classes even across gems (a simple command-click on the method, class or module name). You can quickly traverse source trees without the cumbersome grep/find that you'd do in a basic text editor.
Running specific tests is also a snap; just place your cursor in the desired spec and command-shift-r (built in support for spork).
I'm not in any affiliated with JetBrains (swear); just a satisfied customer.
urgh... yes I've used it but I was hoping for something a little less 1980's. We can have multiple windows now ;-)
I find it useful to be able to see the current stack, drill into variables, view my code with syntax highlighting, and the current line all at the same time. It helps me understand where I am.
RubyMine just tweeted: "We welcome all NetBeans users to start evaluating RubyMine as your new Ruby/Rails IDE! Expect some great news very soon on our pricing page!"
It is a good product and superior to NetBeans in many ways. The price was a little on the higher side though (but worth it anyways). If it comes down, I expect RubyMine to become the de-facto IDE for Ruby/Rails.
Yep, I think consolidation in the Ruby IDE space will be a good thing. All the super elite people are using VIM / Emacs, and all the lazy people like me are still using Textmate. If RubyMine can develop a decent size community around it (with plugins, snippets, colors, etc), I'd definitely use it.
They really have to speed it up though. Feels like slogging through mud when compared to TextMate. Not sure if its possible given their java codebase.
For open source projects it is "free" and $69 for a one developer license. That isn't too bad. I use PyCharm and it is pretty good and the Jetbrains team is always responsive to my questions.
I've only found NetBeans useful under Windows. I've had a number of interface problems on Linux and OSX that make it unusable. (Dialogs showing up blank all the time and other problems)
I've found netbeans to be faster on windows, certainly. Not had any of the issues you mention on OSX - just general 'java is slow' issues (especially when compared to same version of nb on windows).
On Linux, I got tired of Eclipse-based Java apps not working properly - blank windows, as you mentioned being a primary issue.
I have used NetBeans both on Windows as well as on Ubuntu. I haven't noticed slowness on Ubuntu (with 4GB RAM) apart from the loading time. I have switched to VIM since then and haven't used NetBeans in the recent months though.
Is anybody actually using NetBeans for Ruby? I tried it very early in my exploration of Ruby, but quickly switched to gedit, and eventually emacs (using technomancy's emacs starter kit). Now that I use emacs, any other editor (particularly one that requires a hand on the mouse) feels like an RSI producing nightmare.
Netbeans has some very nice autocomplete, drop-down help, and refactoring and debugging stuff for Ruby, but I could never use it for serous code editing. I liked it for the WYSIWYG GUI editor when building some Ruby GUI apps, but much prefer vim for general coding (vim plugins for Netbeans notwithstanding).
Honestly, I feel like the title here is a little misleading.
The message from the netbeans team is that they just didn't have enough resources to support both Rails and upcoming Java 7 and so they had to drop development. They also are encouraging the community to pick up and continue development of Rails support for the IDE.
I guess it's possible that a lack of resources is just an excuse to remove that support, but to me this seems like a genuine reason. I don't get the sense that Oracle just pulled the plug for political reasons.
I use Netbeans for java and used it when I started working with Ruby. But I switched to RubyMine and really enjoy it. It took me a while to convince myself to pay for an IDE after previously using Netbeans and Eclipse, but the cost isn't high and it has really made me speed up my ruby work.
This morning I went looking for a new IDE to replace Netbeans/Aptana and found VIM + Janus (https://github.com/carlhuda/janus). Janus is a collection of customizations and plugins that make VIM look & feel almost like a GUI IDE.
Absolutely agree. I also highly recommend FuzzyFinder if you like TextMate's "Go To File" (CTRL+SHIFT+R Eclipse, CTRL+SHIFT+O NetBeans), and also combined with the Project plugin.
The tutorial uses Project (I would say complementary more than an alternative to NERDTree). It depends on your needs, and I use both. NERDTree for general navigation, but the way Project buckets content results in some flexibility as well. Along with using netrw to do remote development via scp/rsync, you could also create a hybridized project of local and remote content and run a vimgrep on all files. There are at least a few applications for this with, maybe a quick way to manage config or view log files (if they're small enough) on multiple servers/shards for example.
Additionally you can also have it run scripts to switch dev environments when you switch projects. Project switches are triggered when you open a buffer for a file in a different project. If you are working on several platforms and have multiple build scripts, I find this to be helpful. One limitation though, if you have a split or tab with multiple buffers from different projects, the most recent opened project's environment persists.
So FF + Project's workflow is: (1) keys (,f for me) to "Go To File", (2) type in a search pattern, (3) open a file with your environment fully setup and ready to go. For me it has been pretty good so far over the past year or so with a few minor issues.
I use this combination as well with one addition - topfunky's peepopen (http://peepcode.com/products/peepopen). Works like an improved version of Textmate's fuzzy find - even displays Git metadata inline. Its not free - but its worth it imo to have something a little more polished.
Netbeans is open source. The plugin is open source. Oracle is simply trying to cut projects where they have no direct or indirect benefit. I'm sure they would be happy if someone else maintained the plugin. Netbeans has a growing community of plugin developers. There is a community Clojure plugin, for example:
Sort of sad to see this happen, but I'd prefer they keep netbeans going and streamline it to be really good at a few things, rather than being 'ok' at lots of stuff.
Ruby doesn't really fit with Oracle's offerings, although PHP doesn't either. Perhaps PHP support will be dropped in the future?
Ugh, I hope not. I've wedded Netbeans to my PHP development process (and projects) rather tightly so the thought of having to switch (especially to something I find inferior like Eclipse) makes me nervous.
I agree that Eclipse isn't the ideal IDE for PHP (or anything that's not java), but I don't mind it too much. Maybe it's just because I've gotten used to forcing it to do what I want.
I would love to know why you think netbeans is inferior to eclipse? Perhaps for Java, it's a different story, because Eclipse is so much better than NB.
The number one reason would be debugging. The debugging environment as it relates to PHP with Netbeans is simpler to configure and easier to use than with Eclipse. There are other reasons but I feel they are more subjective than the debugging issue. I've tried Eclipse many times over the years and have always either gone back to the old Zend IDE (before they moved to the Eclipse platform) or Netbeans (lately) in frustration.
To be honest, I frequently use Vim for quick editing but when debugging PHP Netbeans beats Eclipse hands down.
RadRails and Redmine are going to be the closest to the Netbeans Ruby support, others have mentioned those already. Something closer to the heart of many Ruby developers is Redcar. It's a text editor installed as a Ruby gem generally meant to replicate Textmate functionality.
I'd love to see a cross platform variant, influenced by TextMate's design. I tried out redcar last year but it wasn't ready for primetime then. I may be confusing redcar with one of many other editors I tried out last year, I'll give it another look.
update it was redcar that I used last year, and it has improved quickly (kudos to the project team). That's a great sign. Kinda psyched to have something textmate-ish on ubuntu! The themes are pretty hot, and the bundles are chock full of languages.
This is the killer feature of tools like emacs and vi: Your vendor cannot abandon you and the code has a well-proven capability to be ported to just about any platform.
The cross-platform aspects of these tools is way undersold. I have all my config checked in at BitBucket. I can sit down at a new machine, check out my emacs configuration, tweak a couple variables if things are non-standard on the box (paths and the like), and I've got my environment completely set up, whether Mac, Windows, Linux, or something else.
I completely agree with this. Emacs is an amazing environment that can be customized and extended in so many ways. It also has support for just about anything you'd ever want to do.
It's also nice that you don't have to have N dev environments installed to support each language you want to play with. I work in Java, C++, Python, Erlang, HTML, JavaScript, and Ruby fairly regularly. Installing and learning a completely new environment for each would be obnoxious.
I used it for a couple of days, then switched back to TextMate. I am going to watch the PeepCode Vim series (http://peepcode.com/products/smash-into-vim-i) and I think that will give me some momentum.
Mhh, it takes some time to get accustomed to vim but it's worth it. Just don't give up too early.
I'm using XCode and vim ... XCode for iOS development and for everything else vim. Prior to that I used TextMate ...
After a few weeks of constant vim usage it happens that I start typing vim commands into XCode (or into any other bigger text field). For example I regularly press a/i/o before I start typing the text. Or I try to block mark/yank text selections, etc. instead of use copy and paste :)
After a few weeks of regular usage you start getting a grasp on vim and will constantly ask yourself why all the other editors suck :)
(I tend to use console vim - not macvim. Macvim somehow doesn't feel right.)
I hate to lead this further off-topic, but I'm trying to get into Vim myself -- how to you approach having multiple files open? Buffers? Tabs? Windows? The netbeans/visual studio/et. al. tabbed approach has always worked well for me, and I'm finding that to be my biggest stumbling block in switching to Vim.
Try the MiniBufExplorer plugin. It gives you a list of all open files in a different buffer, and you can switch to any of them by just moving into that buffer and pressing enter while on the filename.
I use splits and vertical splits. Tabs are an option too, but I usually don't leave that many files open at once.
For browsing files inside vim I use NERDTree (or if I remember the filename :e <filename>). To quickly open the file's headerfile I use http://www.vim.org/scripts/script.php?script_id=31 - it takes only a :AS to open the file's header file (vice versa) in a new split. (Hint: Map a key to next split/window to switch fast between them. I remapped F1 for this purpose.)
If I really need to have many files open I just open a new tab/window in my terminal emulator. (There are also tabs in vim but I had not a chance to get used to the tab commands yet. gvim/macvim offer more abstract tab support - the tabs are real vim tabs but they are represented as UI tabs. So you can still yank/paste between them but have a more Visual Studio'ish feeling.)
If you're on Mac OS X or can run a VM, TextMate is a wonderful project editor. It was the reason I switched over to Mac for development. It does everything I want for development, and gets out of the way for everything I don't want. It'll take a little to get used to.
Autoformatting code is controlled under bundles (format all, re-indent text).
Go to definition/declaration is functionally done with go to symbol.
There's support for code completion but I haven't used it much.
I'm wondering how come Oracle hasn't kill NetBeans yet? Considering the effort they've put into Eclipse and the aholes they are this seems a logical move.
They've been bought by AppCelerator so I highly doubt their longterm focus is going to be on Ruby. Will probably mainly be Javascript (to support Titanium Mobile)
I'm sure there's much more to it, but my initial reaction was to think it was oracle snubbing salesforce in their acquisition of heroku. In the enterprise (which sf targets) oracle is revered, and a robust oracle IDE does help give credibility to ruby/rails. Aptana and ruby mine are the only other ruby/rails IDE's that your run of the mill IT manager would consider full-featured im aware of, and neither of them are backed by a brand as respected in the enterprise as oracle.
This is an unfortunate move, though I expect that the community will do what it can to pick up the pieces here (which Oracle seems to state openly that they want to see happen). NetBeans has improved so much over the years, I hate to see "official" features like this get tossed out.
"Second: Although our Ruby support has historically been well received, based on existing low usage trends we are unable to justify the continued allocation of resources to support the feature."
That is really all they needed to point out (for me).
NetBeans is open source. At least OpenOffice and Hudson have already been forked. Are there any obstacles (name has to be changed of course) for forking this project as well?
>>However, we strongly encourage our community of NetBeans Ruby users and developers to volunteer to take on development of Ruby on Rails support for the NetBeans IDE.
I started out with textmate way back when. I actually started using NetBeans in hope that it would have decent support for Actionscript 3. I happened to find out about the Rails support later on & was very happy I did.
Here's some of the reasons I switched:
1) inline syntax checking
2) ability to click through to called method (command + click in os x)
3) ability to have multiple views of the same file (clone window)
4) ability to have split screen editing.
5) debugging
6) inline documentation instantly available for any built in, plugin or internal class (ctrl + space).
7) code completion / suggestion (this is not a lazy thing -- it's helpful for discovering what's available in any given class)
8) indenting in textmate was never fun. In NetBeans just command + shift + f and the whole file is properly indented.
Those are some of my reasons. Hopefully there's something else out there that can replace NetBeans for me. I've been extremely grateful for it -- especially at its price.
I would bet a lot of them started as Java developers (who generally must have an IDE) who were experimenting with Ruby and used the tool they had in hand. There are also probably a chunk who want a "real" debugger.
And it had RVM-like features before RVM became popular ... you could specify what version of Ruby and what server (mongrel, web brick or thin ... if you knew what you were doing) you wanted to run your app against
Well, it's probably just me but Netbeans support for Ruby was as good as Eclipse(PyDev) support for Python i.e it kinda worked but it left a little to be desired, I for one am not gonna miss it.
This is actually good news. The company I work at would never hire a Rails developer that uses ANY sort of IDE. Learn to use a good text editor and learn the Rails CLI.
What about those who know a good text editor, know the Rails CLI, and still prefer an IDE? There are worthwhile features that an IDE can do that a text editor can't (easily).
If I were running your company, I'd focus more on the individual I was hiring and what they bring to the table, and less on if they're using the hippest tools.
