Hello, I have been looking at 2 possible solutions for moring our ERP system into a web based enviornment and the poros and cons of each. I have included a short overview of each option and would appreciate and feedback and recomendations you may have.
This is one option, to set up a WebSpeed server and build our app in WebSpeed. qwe have lots of Progress expereience here but new web development skill will have to be tought in order to build a user friendly application.
Another option is to have the Progress developers build a backend to run on our AppServer. We also have a .NET Developer who has more expereince doing web apps and web apps. He could simply use teh Progress Framework to make calls out to the AppServer to proccess all the bussiness Logic.
The first method seems to be a but quicker development, but the UI and Bussiness Logic are one in the same. Wher as in the second method the development would take longer but the UI and Bussiness Logic will be seperated and this will provide significanly more custamibility on the front end.
Consulting in Model-Based Development, Transformation, and Object-Oriented Best Practice http://www.cintegrity.com
I guess i can clarify a bit. I come from an OO background and my coworkers all come form many years of Progress. Now the current progress app we use is going to a SQL backend, and we want to stay with a progress backend. This will require the development of our own ERP system, witch we are fully capable of doing. The backend ERP will be in progress running on a AppServer. The froentend is what I am reserching and I have come down to either a WebSpeed frontend coded in Progress/HTML of a .NET Frontend coded in C#. Now i have more experience with the C# but I wasnt to give WebSpeed consideration as well. I have having a lot of problems finding much information about WebSpeed though, beyoned the documentation, and no luck finding any samples. I also can't sigure out if the WebSpeed seerver is a core component of Progress, or an extra that would would need to purchase.
And my point is that there are more choices than that and it should be business issues which drives the choice.
I have to agree with everything Thomas wrote, but have a couple of things to add. It seems to me that your biggest question is what language the front end should be coded in.
As already suggested, the front end should be so well separated from the backend, that in 6 months or 2 years or 10 years, you can simply swap the existing one out for a new one. There is heaps of materials here on PSDN that will help you towards architecting such a system. Search of the OERA (OpenEdge Reference Architecture) white paper series. Also, be sure to download the Reference Implentation application AutoEdge.
As for the front end, you need to consider things like: who is the end-user; what client operating systems do they run; can you install software on their machines; are there going to be firewall or other security structure and how will you get around them; is an HTML application (perhaps with AJAX) okay or will they need a "full GUI client".
These days, there's a lot to be said for a well designed HTML + AJAX application, though don't be fooled into thinking that HTML + AJAX is easy. You'll need someone who really knows what they're doing.
There are certainly advantages to coding business logic and UI in the same langauge, but if there are more than just 2 or 3 developers they'll probably be so focused on their particular part of the application that they'll no want to dig into what the other side is doing anyway.
BTW, if you're looking into C#.Net because you'd like to be able to use all those cool .Net objects out there, you should definitely consider joining the 10.2A Beta program as 10.2A will allow you to use those .Net objects in an ABL UI while still programming everything in the ABL that we all know and love.
I think he is already committed to a zero footprint web client and the only question seems to be whether to get .NET involved in the delivery. I don't see any plus in using .NET on the server, especially since it limits your choice of servers. But, there are still more choices than two.
While, yes. I guess my original post wasn't my real issue. We need to move to a compleatly web based solution, as we have a variety of clients systems on and off site. switching to a web based system will significantly reduce the problems we have with client based software. We also want to move to a setup where the frontend and backend are independed. With this in mind we looking to move to the Progress Backend .Net Frontend. the problem I am running into though is with WebSpeed and how to do the frontend, mainly the advantages and dissavantages of .Net and WebSpeed. We have experience in .Net and Progress but no WebSpeed experience, and I have been having problems finding information and demo's of WebSpeed. beyond that some of the WebSpeed based sites i tried to look up are now running .Net. As far as I can tell it is simply becouse .Net provides a much richer UI development envieronment then the other.
Not wanting to put fuel on the fire, but has anything changed in terms of agent licenses in this type of arrangement?
We have had a gradual move away from PSC as a result of moving down this path. The team found it "interesting" moving between the two development environments - just the small conceptual differences in development style, more than what you would see in say a C# Java environment. Added to that was the cost confusions in serving 50 known users and NNN faceless users.
Changed since when. OE10 has named users, registered clients, and access agents for license by person, license by device, and license by number of processes needed to support unknown and casual users.
A great option is to use Adobe Flex + Progress.
Some examples textbooks.