The Mudcat Café TM
Thread #91460   Message #1745439
Posted By: Mark Clark
22-May-06 - 12:06 PM
Thread Name: Tech: Digital Tradition Programmer Needed
Subject: RE: Tech: Digital Tradition Programmer Needed
In my experience (35+ years), software developers will usually favor solutions with which they are already familiar. They also gravitate towards solutions that simplify the work of development and maintenance, often without regard for the end user of the software. Once developers have had significant positive experience with particular technologies—database, language, platform, etc.—that technology becomes the basis for all their modeling thought processes.

My observation has been that Dick and Susan have many years of thought and experience behind them and maintain a solid understanding of the ways in which the DT is used and the search styles that best support the intended audience. I can see no advantage to shoehorning the DT into a relational database paradigm except to provide developers with a more familiar tool set.

High speed full text indexing and searching is a mature technology. NeXT Computers (NeXTSTEP OS) in 1990 came with something called the Digital Librarian. This automatically maintained full text indexes for every file on the system and allowed the user to find any file in the file system instantly on the basis of content alone. This on a processor operating at 33 MHz, one tenth of today's processor speeds.

NeXT's Digital Librarian became the basis for the Internet search facility known as The Gopher, widely used before the advent of the World Wide Web. It's interesting to note that Tim Berners-Lee used NeXT computers to develop his original World Wide Web servers and browsers. Recent Microsoft operating systems also come with a full content indexing option.

I have no experience with AskSam technology but, reading their site and seeing how it's used, I'm guessing that the developers have incorporated the most advanced full text indexing and search algorithms available. I don't think I'd dismiss AskSam as outdated technology.

I think the key for DT development and growth is to isolate the DT on a dedicated WinTel server hosting only the DT and AskSam. There would be no direct user access to the server. Rather a network communication protocal known as Web Services would be employed to give client systems (which can be other servers or end users' systems) the ability to maintain and query the DT without placing any load on the local system. The DT server, then, can be sized to provide the desired level of performance based on the demand.

This solution also means that the Internet DT "service" will have searching capabilities identical to the version distributed on CD-ROM. Web sites such as Mudcat would use Web Service technology to submit queries to the DT but this will be invisible to the Mudcat user. There need be no change in Mudcat's apperance, user interface or performance. In fact, the DT Web Service could be physically located inside Max's installation if that is deemed best. I tend to think an externally hosted solution would offer greater power and flexibility at lower cost but I probably don't have all the details needed to say that with absolute certainty.

      - Mark