Gossamer Forum
Home : Products : DBMan : Discussions :

New project, open to suggestions

Quote Reply
New project, open to suggestions
I am considering doing a DBman customization for another application (number 3 for me) and have enough lead time to consider the general aspects of the setup and to take suggestions. The end result is a database from which I create static hyperlinked indexes to PDF files. These indexes and PDFs are distributed on CD as an annual Research Report. In the past I have built the database from lists of keywords contained in wordprocessor files that are sent to me by the contributing authors. Last year we had 192 separate reports to process, so with student-worker assistance the job was manageable. So why fix it if it's not broken? As I said I am open to suggestions and I enjoy learning. I have read many DBman FAQs and installed several MODs to create my first and second functioning DBman installations, but by no means have I read them all nor attempted some MODs after reading the instructions.

Thanks for a very informational Forum and the great MODs.

A description of the project follows as well as my current thoughts for structure of the possible DBman installation.

Annual Research Report including about 200 submissions from 30 institutions (of higher learning).

Each report has 1 submitting institution, multiple authors at each institution, 1 title, 1 statement of objective, 1-4 authors (may be at different institutions), 1-10 crops, 1-10 weeds, 2-20 weed control treatments. These are the keywords and they adequately describe the content of the report and ideally the end user should be able to search by multiple keywords. Currently the end user selects from one list (author, crop, weed etc.) much like an index in a book.

The Research Report on CD is produced on a tight schedule as crops are harvested in Sept-Oct., and harvest data must be analyzed and written-up in Nov. for distribution in Dec. There are last minute submissions and changes, but we deal with it. Ideally contributing authors would submit their keywords via internet forms, the problem is we insist on accuracy and expect all authors and 2 addtional reviewers (up to 6 persons) to approve the list of keywords. The procedure last year was to have 1 author prepare and print a wordprocessor file for each report, the printed copy was approved by all reviewers, and the file was emailed. It may be too much to ask to have all 6 reviewers huddled around a computer and agreeing to click the submit button. I prefer to let them take a printed copy for review at their convenience and perhaps one author then submits the keywords.

Regarding the nature of a DBman file setup as I see it:

I could see a main database that holds (for each of 200 reports) the ID (db_key), insitution, title, objective, and perhaps lead author or contact person, these fields to be completed by authors. And following publication; Volume, Section, page number, and link to PDF file, these field assigned by administrator (us/me).

Since, for each report there are multiple crops, weeds and treatments, each of these would be a seperate database, linkages to main database via page number or db_key. Thus I can display records from all databases that have the same ID number. This would be the reviewers copy. Sounds like the relational MOD might be useful. Sounds like something I can do.

After all submissions are in, we publish to CD, there are no more submissions, so the database does not change. Now the users are persons needing to find reports that include some combination weeds, and crops (or any other keyword). I am uncertain if the setup of databases described above is the best for this multiple field search in which the results must include all the search terms specified?

The next question deals with doing the multiple field search using the CD, but without an internet connection. Adobe Acrobat has such a function, as does Wilbur, but both require the user to type the keywords, I would prefer to present a drop down list or checkboxes for desired combinations. It could be a series of questions, such as, pick a crop, submit, possible weeds are presented, pick a weed, submit, possible treatments are presented, select multiple treatments, submit, reports that include selected treatments for that weed for that crop are listed and linked.

Sorry this is such a long post, please feel free to comment on any portion, do not feel you need to present a complete answer. If you want to help and want more information, I can provide links to last years indexes as it would give you an idea of the limited scope and provide a visual example.

Thanks again in advance.

Joe Matthews