Hello All,

At work we run on system-i with a DB2 database and a highly custom application for our order entry program. That program is all green screen/telnet used by our c.s. staff - which actually is fast when you get used to it. Written in RPG/CL.

We want to redo our online shopping cart which runs on an antiquated, heavy, slow, custom web server(Strategi) that talks w/ RPG to bring our order system to the www. We have Zend Server for IBM-i installed parsing PHP applications but my question is: Are we better off building a shopping cart ground up or is there benefit in a Open Source package and marrying it to our system?

My concern with open source is the time it would take to get the two talking but at the same time getting a shop up from scratch seems just as time consuming and you lose all the benefits of testing, security, etc...

Does anyone have any thoughts or experience putting an open source app. on an existing DB similar to this situation? We've put SUGAR on a windows box and used Pentaho with a JDBC driver to get that Open Source app. talking w/ our DB2 but that is day old data replicated in a nightly run. This would have to be real-time transactions against inventory qty's.

I appreciate any insight..
John

    If the cart is just a "front end" and all administration of inventory and system setting is accomplished via existing tools (telnet/RPG or whatnot), then you might come out OK. If you have to write a control panel and so on ... I'd think that'd be a HUGE project, and unless you're outsourcing to someplace with zero COL (Pluto?) I'd think you'd end up way in the hole. And I didn't even mention the payment system ... a huge investment in time to get it right and test, and so on.

    That said, I've generally tried to migrate data to something compatible with the software instead of altering the software to match the RDBMS. But doesn't PHP have a db2 driver? [edit:] and obviously you've stated you're not planning to move data ;-)[/edit]

    As for F/OSS carts, there are some decent ones out there, but it seems to me that docs and/or tech support tends to lag behind professional offerings. Of course, some of those aren't real stellar with "support" either....

      One more thing, though; what we fight with an OOB product is that our need for customization generally exceeds the SW's ability to change via its configuration ... so $n months/years down the road we end up with something that has so many knobs attached that we can't actually upgrade to the latest version of whatever it was we started with.

      So, thinking longer term ... if you expect your business to last a long time and your site to grow and your need for custom additions isn't met by a robust and modular extensible API ... you may end up with a lotta custom stuff anyways.

      Business is such a crapshoot, isn't it? 😉

        Write a Reply...