Friday, 11 July 2008

Desupported Oracle Forms vs JInitiator vs Sun JVMs and browsers

(Please note as it's been pointed out to me that some of the following post was confusing, this post has been edited from the original to make it more clear. One of these days I've got to get myself an editor ;)

For whatever reasons unfortunately some sites don't have the ability to keep their Oracle Forms within the Oracle officially supported versions. As an example you might be a site still running under Forms 10gR1 (9.0.4), a desupported version of Forms (1ogR2 being the most recent supported version when this post was written).

While the recommended approach is to keep up with supported versions, this isn't always achievable in a timely manner, or at all at your organiation.

This presents a number of catch 22s for organisations, one of which I'd like to highlight.

One such catch is where organisations would like to move to the latest browsers on their common desktop emvironments. For example many organisations are pushing to IE7 over IE6 for the perceived security benefits and user preferences of the new browser version. However the last documentation coming out of Oracle for the organisation's specific version of Forms (say 10gR1) only lists certification for older JInitiator or JVM versions, and older browsers (such as IE6 or FF1.2).

What to do? Does this mean you can't upgrade to IE7 or the Sun JVM? Are you stuck with IE6 and an earlier JInitiator version because the documentation says so?

I posted just such a question on the OTN Forms forums the other day and there were a number of useful replies that readers might be interested in, particular from Wilfred van der Deijl and Grant Ronald. Wilfred's reply is well worth the read. Grant also linked back to an earlier blog post of his where Oracle is going with JInitiator and the Sun JVM. Oracle is making no secret of their future choice of JVMs and it's nice to see we're moving to one standarized JVM.

This also brings up another issue, that sites should be careful to base their decisions on outdated Oracle documentation. When Oracle has desupported 10gR1 and are no longer updating the Client Platform Support docs, this means you're likely to get yourself into trouble basing your decisions on such out-of-date documentation. You could mistakenly conclude from the 10gR1 documentation that you must remain on IE6 because Oracle never certified 10gR1 against IE7 and the relating later JInitiator and Sun JVM versions. Yet there's nothing to stop you attempting to upgrade to the latest browsers/JInitiator/Sun JVM, just make sure you test it! Just because Oracle didn't certify it, doesn't mean it doesn't work! Obviously you're already running desupported Oracle Forms versions, so you don't care if Oracle certifies against a specific future browser/Jinitiator/JVM. You just care that it works..... (though be mindful that if you do intend to one day upgrade Forms, try and pick certified browsers/JVMs for future Forms versions to save you unnecessary headaches later).

(To reiterate, I hope readers understand the emphasis here, I'm not saying that the certifications listed in the existing 10gR1 doc are wrong, that's what Oracle did indeed certify against. What I'm saying is as Oracle never certified later versions, assuming therefore Oracle Forms wont work with later browsers/JInitiator/Sun JVMs, is basing your decision on a poor assumption. Basing any decision on a poor assumption is going to give you the wrong outcome to your own detriment. QED)

I can even see in the future this becoming a bigger bind for organisations that want to upgrade to the latest version of Oracle Apps, browsers, the Sun JVM etc, but then discovering they have internal custom Forms systems that haven't been upgraded. Again the assumption you'd need to stay back on IE6 and JInitiator would mean you couldn't move forward elsewhere. Realising you can update JInitiator or to the Sun JVM, and browsers, as long as you test, can be the liberation that your organisations needs.

And personally because I'm interested in JDeveloper, if you remain on older browsers, this brings up a problem for the JDeveloper camp, as Oracle JDeveloper 11g want IE7 or Firefox 2 or above. Imagine having your organisation telling you can't use JDev 11g and all its fantastic AJAX/RIA features because they're running IE6 to support a desupported versions of Forms! I'm sure there is irony in there somewhere!

Of course this could all be fixed by organisations keeping on the supported upgrade path from Oracle, something which Grant Ronald and others have previously emphasised strongly. Hopefully the above helps managers and technical leads understand why letting your organisation lag in supported versions can be such a hindrance elsewhere to the organisation in the future and present, besides just the perceived need to stay with the supported versions. Its not really just about keeping up with support, its about the decisions on the information at hand you make around the software you have at hand, and how up to date that information is.


Mark said...


Interesting post. I’m also curious about using Oracle Forms based on Headstart in combination with the JVM. We had some trouble with the scroll functionality of 10gR2, therefore we stayed on the JInitiator. When the JInitiator will disappear then we have to work through Headstart, since we can’t reproduced the problems in a Form without Headstart.



Chris Muir said...

Hi Mark

Thanks for the comment.

Ah, good old Headstart. People are still using it!

Actually from my testing of the Sun JVM (v1.6.0_07) and IE7 I'm finding the JVM very unstable (I note the 10gR2 documentation certifies IE7 against the v1.6.0_04 Sun JVM, not 1.6.0_07). My feeling is, given the importance of your app, it would be prudent to start testing 10gR2 with the certified Sun JVM and Headstart, and push bugs found back on Oracle to resolve, such that when you finally do move, hopefully the later ceritified stuff has fixed the issues, otherwise you're going to be stuck.

By using Oracle in this fashion, Oracle is going to have much more manpower/skills to resolve the issue, and also clout with Sun to fix the issues in future JVM versions. You'll just need a lot of persistence!

...I guess another reason we pay Oracle for support.