Re: Targeting "enterprise architecture"

The practice I have employed with the adoption of technology is to see how much noise is being made on email lists and forums. Every technology will have issues, but how fast are the issues addressed. How active are the forums? Are people developing similar applications to what my needs are? How quick do new releases come out? Do they wreck a lot with each release (i.e. how painful are releases - I recently got bit by Apache MyFaces)?

EGL is very intriguing as IBM's intentions are good and they *have* taken time to make it work with RPG. But in the end it is somewhat out of their hands as to the success of it, and it will only be around as long as it is making money (the only reason RPG still is around IMO - IBM can still make money from it). If it doesn't catch on and it doesn't make money then they *could* be dropping it simply because they wouldn't be pissing off too many customers at that point (because adoption was so low).
They jumped into a boiling kettle that was already seething with an abundance of fish and are trying to compete in the exact same space. EGL is also "young" as it relates to the technologies it is trying to incorporate (which many are open spec) and that means there is going to be some growing pains in the next couple of years, and that translates into some not so fun coding for adopters and upgrades will probably be more on the drastic side. The good is that they will probably be making good drastic decisions, the bad will be that it has the potential to totally mess with your existing approach to using the tooling/language. Note that this isn't that much different from where a lot of the other tooling vendors are at IMO (i.e. Microsoft and Sun and Adobe).

In my opinion you should take a long hard look at what you and your developers are currently versed in and do your best to make a system work from those knowledge sets. For example, if you have a lot of .NET developers it may behoove you to just stay in the Microsoft camp. In the same breath if you have RPG developers then I wouldn't try to jump ship to .NET or Java or EGL but instead find a solution that keeps you where you are most productive (i.e. OS400+DB2+RPG+other thin GUI technology). There is a good thread going on right now in the WEB400-L list that is discussing stuff similar to your need. We are very much interested in writing an approach/framework that insulates for GUI technologies as those are in a huge state of flux these days.

Here's a quote that I thought of that has been swimming around my head today: "At least I know where RPG sucks vs. adopting some other technology that *will* suck eventually and maybe in some not-so-nice-sucky ways."

Those are my thoughts :-)

Aaron Bartell
http://mowyourlawn.com


DeLong, Eric wrote:

Nope, not really....

IMO, there's not much adoption of EGL (it seems), so opinions are
scarce. I'm losing support for this product (within my department) at
an increasing rate, as we can't find much to get us started. Too many
of the tutorials we've found are for EGL versions we no longer use, or
we applied a fix which changed a dialog, or it just doesn't work as
described... We'd love to find someone using it, and IBM is supposed to
be getting back to us with local EGL customers, but so far, we're just
waiting.
I'm getting really frustrated with this whole infrastructure change.
We've reviewed IBM's Modernization Road Atlas (Roadmaps are no longer
comprehensive enough, it seems..) One thing is certain; there is risk
in the selection of technologies that we adopt, and we are ill-equipped
to assess these risks. Will the technologies we target be viable for
the duration of this project? Which technologies meet our needs for
secure, web-services-based applications deployment across IBM i and
Microsoft platforms? What other infrastructure do we need to support
user access to these services-based applications? Portal server?
Someone here suggested we use SharePoint... Does that support portlets
generated by Rdi-SOA? Right now, there are just too many questions...
My hope with EGL was the extensibility of the runtime. As new
technologies come into fashion, EGL should be able to accommodate them.
This would seem to ensure the relevance of this technology far enough
into the future to ease that worry, but the apparent lack of adoption
paints an alarming picture...
If anyone has anything to say about this, please chime in.
Thanks,
Eric DeLong