<! date> Date: Sun, 25 Mar 2001 19:43:30 -0500
|Subject:||A view of bootstraping with OHS|
Thank you , Henry.
Your comments are always of great interest and value to me.
When I compare and contrast jpOHS to Doug's writing at....
I note that Doug's early interest lies in the direction of getting HyperScope up and running first. That's the engine that would include a grove engine and link database, not unlike BrowseUp, which, as I have said before, appears to be a well-made instance of this class of Web tool.
Well, I'm really on a slightly different vector, one that I hope will ultimately intersect with HyperScope. The direction I am going lies in content production, not in content browsing.
I am of the opinion, and I think my opinion is clear in my earlier post, that no matter what vector you are motoring along, there will ultimately be some kind of API that is the glue that binds all vectors.
My experiments with jpOHS continue to show me just how weak my ability to visualize implementations really is. That's why I jump off bridges and build implementations.
For instance, my latest jpOHS paper, at....
.....tells the story that I ran into all sorts of problems when attempting to combine both synchronous (blocking) transfers and asynchronous (non-blocking) transfers. Over this weekend, the April Java-Pro came with a paper on using SOAP to solve the very problem I was facing. That program wanted me to boot Tomcat, the servlet engine. Well, guess what? Darned if good old uncle Bill hasn't tossed yet another mess out there for me and others to solve: Windows ME loves to tell me it's out of environment space when setting classpath variables in the batch file that runs just fine on NT4 and 2000. You gotta love jerks that continually toss barriers into the path of java hackers.
What came out of that? Simple. Find some way to do SOAP-like things without running SOAP. I think I'll have that running sometime next week.
<! close> Sincerely,