From: Jon Paris
EGL next?
Next for the chop? Next to be ignored by users?
My money is on the latter.
I'm not sure how to address the points made here, but let's see what we can
do.
Until IBM either makes WAS so small, fast and easy to use that it requires
as little attention as Apache then anything that relies on it will have a
hard time in the market.
I'm not sure what you mean here. Do you mean WAS must be comparable to
Tomcat, which is a small, lightweight application server, or do you mean WAS
must actually be comparable to Apache?
If you really did mean that WAS must match up with Apache and that somehow
J2EE will have a hard time in the market vs. CGI, then you're dismissing the
gigantic base of Tomcat-based and JBoss-based applications out there along
with WAS. And it's not just my opinion; I think Sun will argue with you
about whether J2EE is a large market <smile>.
So I assume what you really meant WAS vs. Tomcat, since both Tomcat and WAS
use Apache as a front end. In fact, WAS has great features for segregating
traffic between Apache and WAS, so that the static traffic can take
advantage of the unique capabilities of the System i's powered by Apache
HTTP server.
And finally, I guess I don't understand what's so hard about exporting an
EAR file from WDSC and then importing it into a server on the host.
Configuring a WAS instance and getting it running takes less than an hour.
I just wrote an article on installing WAS; it took me a couple of hours.
And that's on the System i, never the simplest environment to install
things. The Zend stuff isn't exactly plug and play.
But an interesting issue would be whether I can export an EGL program to run
on Tomcat. I haven't tried it yet; maybe I'll try later this month. If EGL
applications run on Tomcat, that removes nearly all of your arguments
(unless you insist that CGI is better than J2EE, which is a different
argument and has nothing to do with either WAS or EGL).
EGL is the latest name/version for software that has beena round a long
time.
Well, true, if you also consider i5/OS to be latest name/version of SSP or
CFP (or heaven help us, CCP!). I find it interesting that you would imply
that length of service is some sort of negative, given the fact that we work
on one of the longest-lived platforms ever created.
And just like i5/OS has evolved and expanded from its forebears, EGL is
nothing like its predecessor of some 20 years ago. EGL started life as
Cross Systems Product, went through a major revamp to VisualAge and a
client/server paradigm, and then got a complete facelift to become EGL,
supporting the very latest language and UI capabilities.
Add to that the powerful WYSIWYG capabilities of the Rational tooling, and
it's pretty hard to find another tool as complete. I can create a web page
that dumps a file by dragging a record onto a page and adding a single line
of code ("get customers") to an auto-generated page framework.
In fact, most other languages don't match up very well at all. Heck, PHP is
nearly 15 years old and was originally designed to do home web sites and
very little has happened to the language over the years. Only recently were
some object capabilities grafted onto the language, a la C++.
It has always been good, but it's tendency to ignore System i
specifics and insist on COBOL or Java generation (i.e. no RPG) has made it
an also-ran and I see no9 sign this time will be any different.
This point is true, but that only matters if you plan to use EGL to write
applications. Me, I use EGL to write the user interface and call iSeries
business logic. EGL has excellent capabilities for that particular bit.
Me, I use RPG for my business logic, and I don't use code generators to
write my RPG. No, to me the part that's been missing in the System i is the
ability to expose RPG business logic to the web easily, using WYSIWYG design
tools and with embedded debugging capabilities that we're used to, and EGL
is just the ticket for that.
I'd like to be wrong but ...
We'll have to see!
Joe
midrange.com
