IBM is placing a major emphasis on development tools as part of its strategy. It's a bit of a departure - tools have never been a big business - but one that many other leading software players are aping as well. A new article on internetweek takes a look at how IBM's tool strategy stacks up with some key rivals.
Read Analysis: IBM's Tool Strategy --- How WebSphere Stacks Up
Tools have never been a big business? Maybe not for IBM, but AFAIK Merant, mks, Borland, ... make quite some money selling their tools... okay, not the largest companies, but for me, that is "big business" category...
Can't say I'm a big fan of WebSphere, so I'm not surprised BEA is giving them a run for their money. Perhaps if IBM would quit trying to stack the deck in favor of their propreitary technologies there wouldn't be so much competition from such a relatively small company.
In the tools market, I know IBM has tried to sell us development tools that cost upwards of $5,000 A LICENSE. In general, I don't like using tools I can't also run at home, and I'm not plopping down $5,000 for any piece of software right now. My first car didn't cost that much.
If IBM is doing anything to get myself and some other techheads I know to feel better about Big Blue, it has to be Eclipse and what appears to be a more progressive approach. IBM was not known for being completely up-to-date on Java and there was not place that used to be more apparent than Visual Age.
I think if it can get past inflated price tags, bloated software, and yesterday's technology, they probably do stand a good chance in the tools market. Considering how many companies exist to make tools, I imagine there must be some money to be made.
IBM's "Websphere Studio 5.0" is based upon Eclipse
Websphere Studio pricing:
WebSphere Studio Application Developer V5 — $3,499 per user
WebSphere Studio Enterprise Developer V5 — $7,500 per user
Tools are good but 5k+ per developer is insane.
why would you spend 5k for any tool when there is
IntelliJ and Eclipse.
Those are two of the top tools, and neither is close to 5k.
if you have not seen intellij 3.0 check out the early access program.
Tools are great but 5k+ is insane
This is rather a broad statement isn't it? I mean, if the ROI is there - your developer is going to be >$5k per year more productive, what is the problem?
I think just bleating about the cost of tools shows how naive most developers are - personally I want to be able to build systems that are better architected with superior code and deliver them to my customer faster. If a vendor can prove that by spending the requisite amount of money I can achieve this, then I would be a FOOL to not do it. An absolute fool. And although I don't believe that IBM succeeds in this, there are some companies out there that have some kick-ass technology that really make a difference - it has to be up to the developing organisation to be convinced that it will deliver.
I think the thing that will lead to the ultimate downfall of the Java community is simply this head in the sand, cheap is best attitude.
I think the thing that will lead to the ultimate downfall of the Java community is simply this head in the sand, cheap is best attitude.
I've used all of the expensive tools (TogetherJ, Rational,etc) and zero of them have been worth the cost, to me at least. Hey, whatever tool works the best, right? The Java Community has a best is best attitude, not a cheap is best attitude. And many of us have found that we get far more productivity out of tools that don't cost $5k. If I can be >$5k per year more productive without spending $5k, I'd be an absolute FOOL to do it any other way, as would my clients.
What companies would you suggest are creating kick-ass technology that makes such a difference that organizations ought to go spend a tremendous amount of money for these tools? How would you define "better architected with superior code"?
The reason that it is hard to get excited over IDE's is that they merely generate code faster which helps in about 10-25% of what a developer does.
Having more code to test, debug, document and maintain actually becomes a huge negative for overall productivity and flexibility.
Development/runtime frameworks are the better alternative. Luckily we are starting to see some of these tools coming to market such as Altoweb's Studio and BEA's WebLogic Workshop.
These types of products genuinely raise the level of abstraction and add value across the entire software lifecycle.
Unfortunately, IBM and Oracle solutions are still stuck in code-generation H---
This is rather a broad statement isn't it? I mean, if the
>ROI is there - your developer is going to be >$5k per year
>more productive, what is the problem?
i really dont think that WSAD adds a 5K productive bonus
on intellij. In many areas, there is a productive decreasement when you use WSAD. (i regulary check
the new EAP versions of IDEA...this is really nice software)
But we can expect more in the future, with all these tools for eclipse, companies like borland and intellij have to do excellent work to compete further on.
If the tools sector is a cash cow? Who knows, but its by far one of the most exciting areas... :)
Does Microsoft sell tools at $5K a pop?
Anything that needlessly increases costs just makes .NET look like a more cost effective solution to the money-men. It might be easy for myself as a developer, or my immiediate manager who understands the technology, to justify that kind of cost. Unless of course we see another tool ten times cheaper that has similar functionality, and then the people who hold the purse strings will want to know why we aren't using that instead.
Unlike MS, we don't have to go with any particular company for our tools. $5K a copy is not competitive. We dumped Visual Age in favor of NetBeans/Forte, and later dumped Forte in favor of Eclipse. Cost had nothing to do with it, we just wanted a better tool.
A high price tag does not mean its better.
This thread appears to be skewed towards cost and not value. It's all relative to project size and risk.
The tools market is still pretty fragmented. To me, Eclipse appears to be the platform of choice for vendors such as Rational w/ XDE and TogetherSoft (why WebGain?). BEA is buddying up with JBuilder to get a tight integration story just like Studio and Websphere although the runtime environment and SWT are in IBM's favor. Integration, code generators, refactoring, graphical representation, and aspect-oriented programming are all $$ differentiators around core IDE features. My guess is they will all be standard equipment in Eclipse sooner or later.
if i really get a tool that makes me productive i would pay the earth for it. But will tools sell well on an average- i doubt.
Most development shops are not ready for it.
They have already got burnt by people selling them half-baked tools. Ditto for developers. Nobody wants to buy a tool that is ejb specific either. I think individual developers will pay for tools (price/freelabour/etc). It is a pity but the market has to be mature enough for it and it just isnt now. And it hasnt helped that so many vendors have sold immature tools with shoddy
user experience - they have killed the market anyway
Now that Im at Oracle I would suggest people check out 9iAS and JDeveloper (in fact if you download JDeveloper you get the core J2EE components - OC4J - automatically)
Having used VAJ since 0.9 and taught it in the UK since day 1 (I taught the first VAJ course in the UK) I would argue that for about 3 years it was head and shoulders above the other Java tools and brought a serious OO environment to a not very well designed OO language. Most of the problems routinely reported about VAJ were caused because Java kept changing whereas Smalltalk had been stable for years. Oh and of course lots of newbie java programmers back then had little idea about the art of OO programming (not unreasonably) and what support was needed or desirable or how to drive VAJ. Eclipse is very good (the core of which was built by the same guys who built the core of VisualAge just in case some of the VAJ detractors didnt know) and I assume WSAD will also be good.
However, people wont use WSAD for anything other than websphere so that leaves BEA and Oracle to slug it out for app servers (leaving aside the GUI & c/s market) and Oracle for integrated app server and developer tools (BEA dont have any Java tools of their own of course & for corporate buying that IS significant).
My experience is that programmers like tools the way sales guys like cars (provocative ? no....;-) and that different features appeal to different people for different reasons - Unlike MS there IS choice & thats a good thing IMNSHO.
Seriosusly there are different justifications for different circumstances - for instance a start up based around a small team of expereinced developers would be different to 100 government staff who have never learned Java. For the former they could use anything & do the job, for the latter use of a decent, effective, consistent and supported tool is very significant
IBM accounts rarely buy development tools away from IBM because they get VERY good deals on the list prices. For everyone else there is a choice.
Im glad I ended up at Oracle because having spent time at WebGain apologising for that tool I can at least Evangelise JDeveloper knowing its a damn good product (and at an excellent price)
Biased , no of course not ;-))
peace to all,
Phil McLaughlin (at Oracle just in case you missed that bit ! )
i'm somewhat surprised that the article lacked a comparison with the oracle stack. they may be a little late to the game, but they've got a good lineup:
orion app server: lightweight but powerful; not to mention their acquisition of hp app server; that'll be interesting to watch.
jdeveloper: a solid ide with db (more on this later), app server, and soap server integration.
toplink: an excellent o/r mapping tool. with transparent persistence still in the air, this is not a bad way to go.
oracle 9i db integration: that's another thing that was missing from the stack mentioned in the article, a database server. in order to be a true "one- stop shop", you have to have this.
on an unrelated note...
at the risk of contradicting myself, i've tried jdeveloper, but i like eclipse more. at home, i run eclipse with the genuitec plugin. i think these guys should stop developing proprietary ide's and just do plug-in development. that should help with the whole cost debate.
Earlier this week, IBM announced the availability of Websphere Studio 5.0
The web page says:
"Version 5 of WebSphere Studio Application Developer and WebSphere Studio Enterprise Developer are now available (via PassPort Advantage) for migration, development and testing of applications targeting WebSphere Application Server V4.0.3 and WebSphere Application Server V5.0 (Beta)."
Using my company's Passport Advantage account, I tried downloading
The only version that I can find is the "Early Availability" release.
Is this a beta version?
It appears that IBM's press release was the announcement of a beta version of WSAD 5.0
WSAD 5.0 won't be released before the end of this year - not in August or September or anything else that was promised to us from the sales representatives.
I guess you havent read the announcement, for all existing customers of WSAD v4.0 that have purchased software from IBM through the passport advantage programme and are covered for upgrade protection. You are fully entitled to download and use WSAD v5.0 , this will enable you to prepare you J2EE projects for J2EE 1.3 functionality and get them ready for deployment on WebSphere 5.0 when it is released.
So this is code that is available now !! but only for exisiting IBM customers.
My tooling history,starting later 1996
Java SDK, Command line
VisualCafe 1.8, 2.0
VisualAge for Java 3.x,4.x
WSAD v5.0 and will never look back.
WSAD v5 has been released as Early Availabilty on Passport Advantage. Basically it is GA level Tooling with Beta Level WAS 5.0 runtime test environment. So download it and get coding its great.
It includes both a WAS 4.0.x and 5.0 integrated test environments and allows you to create J2EE 1.2 or 1.3 projects and then run them on the testing environment of choice including Tomcat. this will allow the devleopers to upgrade and still maintain older projects.
Its got all the Eclipse 2.0 stuff in it plus a million other things, like the bit that builds sequence diagrams from profiling data collected from debugging I like that bit, or the improved web designer and visual compisition editor. It also has loads of plugins for data evelopment, JUnit, ANT, SQLJ, XML, WebServices, J2EE etc.
Worth evey penny.
Please dont tell me that Microsoft,Oracle and BEA give their stuff away :-)
When WebSphere 5.0 is available it will be refreshed and available for developers through PartnerWorld. Go and register and get ready its going to kick ass.
its going to kick ass.
I Love it Matt, have you been taking marketing classes :-)
(P.S. I agree WAS 5 and WSAD do kick ass!)
It is quite clear to me that Eclipse/WSAD represent a quantum leap in IDE tools over and above NetBeans/Forte, JBuilder and others, although I admittedly do not have experience of IntelliJ.
My colleagues who use the above tools are mostly switching over to Eclipse for the following reasons:
1. Stability - it seems to be the experience that JBuilder, NetBeans et al. have problems with crashing, corruption etc.
2. Structure - the notion of a Project as defined by the WSAD/Eclipse architecture is far superior to that as defined by JBuilder/NetBeans - more more intuitive and encapsulated.
3. Navigation/Code generation. I think it absurd that someone here has said that code generation may "only" save 10-20% of a developer's time! That is significantly more than $5k of anyone's money. The sheer time saved due to the debugging features in Eclipse (too many to mention), code generation and completion (even guessing what your Javadoc should look like!) and the ability to immediately bring up the callers of a method, implementors of an interface can save a HUGE chunk of one's time - I estimate that these features alone double my productivity (well, they give me more time to surf - ;-).
4. Plugins. A huge amount of work going on here, from XML editors, SQL database plugins, Tomcat, Source-code management and many more.
The ex-JBuilders here have moved on. For myself, who has used VAJ for many years, I miss the ability to drop down the debug stack, change methods on the fly etc etc, but would still probably take Eclipse given the choice.
This advert was brought to you by... me.
I miss the ability to drop down the debug stack, change methods on the fly etc etc, but would still probably take Eclipse given the choice.
Unless I am confused, that still can be done.
Unless I am confused, that still can be done
In VAJ, you could "Drop to selected frame" to a previous stack element at debug-time, effectively re-starting your code from eg. a few method calls back.
Also, whilst debugging, you could decide to change the logic/fix bugs whilst halted in a method. CTRL-S would compile your changes, and drop to the start of the method.
And... you CAN still do it. I have clearly had a total brain-bypass for the last few months and missed the option on the menu! Anyone still using JBuilder should be caught and institutionalised now - you have NO IDEA how much time this feature saves.
And... you CAN still do it
I thought I was doing it (at least partially). I love this feature. I was doing it with Eclipse 1.0 and Java 1.4 but it blew up. I switched to Eclipse 2.0 (beta at the time) and it worked fine.
Users of other platforms - not just IDEs (and you know who you are) have no idea the time this feature saves.
ROI is exactly
why i said 5k is insane.
I doubt you will find a better IDE than intelliJ and the ROI on that is pretty good considering you don't have to take out a loan to get it. and it doesn't cost more than twice as much as the machine it runs on!
Those are list prices. Become a 'partner' and you can get it cheaper. And it is much cheaper with volume licenses. We (my client) should be getting it as as part of an a annual 'contract' without spend any more.
I have been using Eclipse and WSAD. WSAD provides more functionality. If you don't need it then don't buy it. Eclipse + plugins really does a good job for most things. What really is needed is a GUI design plugin for Swing/SWT (Some OSS ones are in the works). WSAD 5.0 comes with a Swing one.
< GUI Plugins >
If they generate code they are insufficient. Such a tool must allow for the separation of view from app-logic. The UI should be externalized. I should never have to look a JBuilder-like generated crap^h^h^hode.
The UI should be treated like an external resource, loaded at runtime, not hard-coded in your app.
It must support full cross-platform UI. You can't do that using the existing layout managers.
Springs&Struts ala Galaxy. I notice that SpringLayout has reappeared in 1.4. You still need tools that provide for the direct-manipulation of the UI. Manually setting up a complex SpringLayout will just as bad as GridBagLayout.
And don't get me started on webapps and their misuse. if Swing was sufficient and Applets worked, we wouldn't be trying to showhorn full-featured apps into a web-browser. HTML, HTTP: Hyper-Text, not GUI.
Mark, did I just meet you at the Driving Range a few days ago, here in SF?
Unemployed Swing Protester
If they generate code they are insufficient.
But is it efficient enough for the productivity gain? Most of the time yes.
>Such a tool must allow for the separation of view from app->logic.
Mine does and has for years. It really isn't that tough. It is all about technique.
>The UI should be externalized.
Separated yes. Externalized - ???
>I should never have to look a JBuilder-like generated >crap^h^h^hode.
Then don't. But I know what you mean.
>The UI should be treated like an external resource, loaded >at runtime, not hard-coded in your app.
It can be done this way. One shouldn't hard code their business logic to one UI. I don't know about an external resource though. It probably is a good thing. When I develop my 'business logic' I try to think of at least two 'UIs' it could be used with (ie. Swing and batch file).
>It must support full cross-platform UI. You can't do that >using the existing layout managers.
Ok. I've never heard of this being a problem.
>And don't get me started on webapps and their misuse. if >Swing was sufficient and Applets worked, we wouldn't be >trying to showhorn full-featured apps into a web-browser. >HTML, HTTP: Hyper-Text, not GUI.
Web apps - I don't build web apps. I do build applications that can be deployed in a web environment. I do have having HMTL as an interface. I only use it (when I get to choose) as a VERY limited UI. Swing needs improvement but it is much better than HTML. BTW, HTTP is not exclusive to HTML. We are using it with Swing clients.
>Mark, did I just meet you at the Driving Range a few days >ago, here in SF?
No. Never been to SF. Been in Santa Clara/Sunnyvale/etc. in early 90's while enslaved with Uncle Sam.
I'm still currently employed. No time to golf. And no money. If I was unemployed - well it would be the same story.
As app servers become commodities, tools like Studio and it's hippy brother Eclipse offer a real way to differentiate J2EE product and still compete against Visual Studio. If you can bring new developers up to speed faster on J2EE technologies and lower development risk - you go girl.
I wonder if the real sleeper of IBM's tools is SWT which could finally offer serious GUI development on Windows and Linux and counter .Net's Windows Forms.
I use IBM WAS v4.03 everyday for work. If you really want to help yourself and the tools that you use to run you apps aren't selected by execs whos total technology experience extends to using e-mail (poorly) look into Orion. It's a great server and exceeds (in my experience) IBM Websphere in every category except for bloat.
What the heck are developers complaining about anyway....you balk at $5000 yet paying you a whopping 100K is good? RIGHT!