I'm still waiting for young Aaron Skywalker to give us GUI for the iSeries.
I think he's brewing something there in Mankato. Do tell Aaron..... :-)
I just need to find the time. Actually if I would just resist about 5 email
thread debates with Joe I could probably put together what I need :-) Oddly
enough, the solution I am working on involves Java - or at least the
prototype does, but the final might not.
I am searching for ways to have what I am working on open source and still
generate money from it (I need to buy diapers - not for me, for my kids.
The wife and I have been pro-creating like rabbits). I would love to build
the base framework, make it open source, and then commercialize an
"enterprise" version of it for sale. The enterprise version would just have
more features, UI controls, etc.
My reason for making it open source is I believe it could change the face of
RPG business application programming on the System i5, and I know there are
many minds out there that could make it light work (and get closer to having
it right the first time).
I have one company that has enlisted us to write an application for them
(just got moved from happening this month to late January), and I will use
portions of that for the beta of the open source prototype. I also have
gotten a good handful of responses from my "solicitation" on the
www.mowyourlawn.com home page.
The concept isn't rocket science, I just kinda stumbled into it with another
project (J2ME on printers) and found that you can make a lightweight
"renderer" out of Java that simply knows how to receive in screen
definitions, paint them, and return actions back to the "mother application"
when predefined events take place (i.e. button click, tab into/out-of field,
keep-alive, etc). The communication between the screen and server program
is done using XML (though I am considering data structures to speed it up).
Though, if left in XML form, the "renderer" on the client can really be
driven from ANY server-side programming language (think about COBOL on
mainframe opportunities). The key is to make the server side easy to code
and create wholesome applications in. That is where the "router"
application comes into play. This app receives in the XML actions from the
client, maps the data to a medium (take your pick) and then dynamically
calls a sub procedure in a program passing event information (using system
API's which have been available since at least V5R1). This part of the
concept isn't original either because I gathered a lot of it from my JSF
interactions (which is one of my favorite things about JSF - mapping an
action right to a function without having to write a lick of control code).
There are other things like "screen stacks" that I have built in so you can
have a similar mindset to traditional 5250 programming. This framework
might sound a lot like programming for the browser - it kinda is exactly
that, except you get all the benefits of web programming (i.e. zero
distribution except for initial client install), and you don't have the
drawbacks of web programming (i.e. different browsers, "back-button" issues,
CSS mishaps, persistence, etc) because the client is entirely written based
on the communities standards and needs.
In the end it all comes down to wanting the framework to be a success. To
do that I will need to spend a good number of hours putting it together. To
do that I inherently need to involve my employer, and my employer gets to
pay my salary - hence the need to make money from it. <grin>
Anybody with insight on how "commercial open source" can be successful? I
read about it a fair amount in SD Times, but don't know how much those
companies fabricate their success (or lack of it). That is the direction I
am proposing this with my employer in hopes we can make it work.
Anybody up for a challenge? I know Joe could add a lot of solid input to
the project based on his experience (and what I have seen of his other
frameworks), and it would create a ton of consulting opportunities for
anybody that likes to bill themselves out. Not to mention the education
opportunities for the likes of Systemideveloper.com, RPGWorld.com and
COMMON. Man, now I am getting excited again!
Aaron Bartell
http://mowyourlawn.com
midrange.com
