Re: what could have been - include HTML in the 5250 data stream

Steve,

The 5250 protocol is stateful, not stateless. (One might say _very_ stateful!)

And yes, in the early days (V3, I think) IBM included a web server called the "Webulator" (or something similar -- I may have the name wrong). This allowed you to visit 5250 screens via a web browser. It also allowed you to code the HTML DDS keyword, and supply HTML tags that went along with the 5250 screens when they were displayed.

What was quickly learned was that the webulator didn't work very well, and adding the HTML keyword didn't really help. The idea never caught on, in large part due to the statefulness of the protocol.

in modern times, with technologies such as AJAX, we're much better able to deal with a 5250 emulator in a browser, because AJAX allows us to overcome the limitations of the stateful to stateless transitions. (For example, Genie from my employer, Profound Logic works great, and uses AJAX for this...)

However, adding more capabilities to 5250 is like putting lipstick on a pig. What would be the advantage of turning PC5250 into a web browser that works over a little-used stateful protocol? IMHO, that's not elegant. It makes much more sense to start converting the IBM i technology to the uber-popular HTTP protocol.


On 4/8/2013 2:36 PM, Steve Richter wrote:

I am interested to know if it would have ever been possible to include
HTML, CSS and Javascript within the 5250 data stream. After all, 5250 and
HTTP are both stateless protocols. Send the data stream down to the client.
Then independently get back a response stream.

I don't know if 5250 has an escape sequence capability. Say a hex'17'
starts the escape sequence. And the 5250 device will ignore all data that
follows until another hex'17'. Or something like that.

The purpose being that regular 5250 terminals would not see the HTML code.
But an emulator running on the PC would be enhanced to process the HTML as
if it were a browser.

And of course the workstation controller and DSPF would have to be able to
handle the response data in the form of the query string and form data.

I am curious if this was ever possible, if IBM ever considered taking 5250
to the next level, so to speak.

-Steve