What Will Be the Next Version of NetBeans?



What is your opinion?

http://www.netbeans.org/servlets/ReadMsg?list=nbusers&msgNo=57150

5.0, I think. At least if I take a look at this and that's not a complete list. There's Matisse, plug-in development support, new CVS, profiler, support for JBoss, Weblogic, Struts, JSF, new code completion, error stripe, editor hints, code templates, background scanning, many new refactorings, new Tools | Options, ant debugger, xml editor improvements, module manager, JSP and HTML palette... and other features I can't remember. Sounds rather like a major release to me.

Update: You can vote here. You already know my recommendation... don't forget that besides engineering reasons there are also marketing reasons.
Comments:

There's so much new stuff, enough for two releases at least, so maybe we should just skip a number and go straight to 6.0... :-)

Posted by Geertjan on srpen 25, 2005 at 01:03 odp. CEST #

And how about NetBeans 2005? Just joking :-)

Posted by Roumen on srpen 25, 2005 at 01:07 odp. CEST #

Actually, I agree with the argument that says that from 3.6 to 4.0 were really fundamental changes, while the difference between 4.1 and the upcoming release are 'only' many new features. So, according to that reasoning, only when there's an architecture change should there be a change to a major release. Or, looking at it another way, only when there's loads of people complaining about changes planned for an upcoming release (such as all the complaints about how people hated NetBeans for deciding to drop the Filesystem in favor of an Ant-driven project system) should there be a major release. And nobody -- not one single person -- has said the following about our plans for the next release: "That's it. I'm dropping NetBeans forever." Ergo, we should stick with 4.2 (or 4.5) but not go to 5.0 (yet).

Posted by Geertjan on srpen 25, 2005 at 02:11 odp. CEST #

I disagree. A major version update doesn't have to be in my opinion only about incompatible changes. Major version upgrade says that there is a big step forward in product's set of functionality. If every such update would be conditioned by incompatible changes, we would not have too much of compatible software... Also, your reasoning is very technical. But choosing a major version has to do a lot about product marketing. With that said, I believe we should go for 5.0.

Posted by Roumen on srpen 25, 2005 at 03:02 odp. CEST #

I disagree as well with the reasoning for major version changes. If that were the case, we would be on Solaris 3 or 4 right now (arbitrary numbers picked). Major revisions are also points of deprecation and I surely hope we won't go through architectural changes very often. The market has been trained for major revisions without major architectural changes.

I am leaning towards NetBeans 4.2 because it has been referred to as 4.2 for a while. If it is going to be called 5.0, call it 5.0 up front. If name changes are going to occur mid-stream, go with the code-name approach and decide the final name closer to release.

BTW, lovin' the work the NetBeans folks are doing.

Posted by John Clingan on srpen 25, 2005 at 03:16 odp. CEST #

John, you have a good point with mid-stream changes. You know, such decisions are not done easily (everybody has an opinion). On the other hand there's still approx. 4 months to release... so it's not that late yet IMO. We'll see.

Posted by Roumen on srpen 25, 2005 at 03:28 odp. CEST #

... and thanks for supporting us. You're also not doing bad with OpenSolaris :-)

Posted by Roumen on srpen 25, 2005 at 03:41 odp. CEST #

It might be a good idea to go straight to 5.0 but: 1. 4.2 has been marketed for quite a while now, so I think it looks silly to change it now. I think the choice for 5.0 should have been made at the moment Matisse was decided to be in the next version. And then it would have been a good choice. 2. Going for 4.2 is not bad either because a lot of folks tend to avoid a x.0 release and will wait for the x.1 release so thy don't get the little bug annoyances.

Posted by Patrick Holthuizen on srpen 27, 2005 at 02:45 dop. CEST #

I think in addition to adding new features, it might be worth bearing in mind that some features will be incomplete. For example, CVS support has been updated but compared to the previous version there are a number of source control solutions that were supported but no longer are (VSS, SVN, ClearCase, etc). To me when I see a new major version number, I have an expectation of great improvements over the previous and if I see regressions I feel disappointed. Obviously, everyone agrees that a number of new features are finding their way into the next release, but there are still regressions in some cases (source control integration, code completion (some might argue), etc) so I think a shift in major version numbers should be done at a point when all the features in the IDE are functionaly complete. My vote goes for 4.2.

Posted by Jesse Beaumont on srpen 29, 2005 at 04:42 dop. CEST #

VSS, SVN and ClearCase still will be fully supported and will be downloadable through Update Center. Code completion should be fixed before release. Do you know of any issue in code completion other than those which are filed which you consider as regression? Thanks.

Posted by Roumen on srpen 29, 2005 at 04:47 dop. CEST #

I would call this version "NetBeans FU" - as in 'F'inally 'U'sable, or also 'f'\*\*\* 'u' eclipse :) (so in that translates to 5.0 in the real world I guess) I love the new features, especially the new CVS - the old stuff was nearly impossible to use.

Posted by karel h on srpen 29, 2005 at 05:19 dop. CEST #

Roumen, that's my point exactly. I know that they will be supported but the current buzz is that SVN might make an alpha release by the time the next NetBeans release is made. I'm not saying that the new features aren't great, I only mean that people who are expecting to see a great new experience when they see 5.0 will see a regression. It doesn't matter that you and others will be saying new profiles will be available from the update center soon, people will still be disappointed. That is why I think when you make the 5.0 release it should be feature complete (ie have no partial features). As far as code completion goes, it's really just the Alt+P thing that some people find annoying. I don't know if it's been filed or not, I've just seen some discussions whereby people find the use of a second keybinding confusing and weird, as well as loosing the javadoc to go with it. Anyway my point is that there are some aspects which won't be fully supported in the next release but will be shortly after, which seems to me to indicate that the next release shouldn't be a major number increment. Just my 0.02.

Posted by Jesse Beaumont on srpen 29, 2005 at 06:17 dop. CEST #

The Alt-p thing is filed and I'm pushing this one actively :-) It's been observed on the usability studies as well. Plus there are other 17 P1s/P2s on code completion that need to be fixed before release. Good points about completeness. So far it seems that community is rather for 4.2.

Posted by Roumen on srpen 29, 2005 at 06:23 dop. CEST #

Post a Comment:
Comments are closed for this entry.
About

Roman Strobl

Search

Archives
« duben 2014
PoÚtStČtSoNe
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
    
       
Today