|
Up and down the land, companies are commanding, "Get the company data on the Web!" Of course, but there's a little matter of having data in four different database systems (not to mention a half dozen other formats in various PC programs). And when you get down to it, there are a lot different kinds of information, each with many gigabytes of data. How are people going to find what they need? Sure, the Internet (or an Intranet) is a great place to disseminate information, but HTML is an awful environment for data management.
All true. The demand for putting data on the Web is real and so are the problems. Is there a silver bullet? Not really, but DataSpot is a Web database publishing program with enough capability to cut through some of the worst problems. Most of all, it offers a way to gather data from many different data sources and process it through a truly powerful indexing and associating engine that makes the information highly accessible to the end user. It lets people search relational databases with the same interface (single field input for search criteria) used by Web search services.
You probably have to do a couple of DataSpot applications before it sinks in how important it is to make design decisions about your data and what the end users may or may not want (or need) to see. This is always good advice, which we tend to follow on an ad hoc basis. But DataSpot's ability to draw in a large amount of data and organize it in so many ways puts the onus on you to think ahead.
The key is to visualize what the end users will see. They will be in a standard browser. They can do searches just by entering a word or phrase, exactly like Web searches they probably know, such as Alta Vista or Yahoo. Once they've started their search, DataSpot begins to open Web pages filled with hypertext jumps based on associations between words and phrases. (See Figure 1.) Users can move further by taking these jumps or by continuing to enter subsearches. (Administrators predefine these associations.) This is how you can open up vast quantities of corporate data to searches that even people who've never used a database can manage. On the other hand, although you're not conducting a guided tour, it's important to design the avenues of search and -- in the case of some kinds of sensitive information -- to limit or control the user. You can do this with DataSpot, but like I said, it takes some planning.
If you're familiar with a SQL view, then you can quickly understand the first part of the DataSpot approach. In essence, you use the DataSpot Publisher to tell DataSpot what tables you want to use, you specify the fields and their properties (see Figure 2, page 36), and it collects your definition in what it calls a DataSpot Web View. This becomes the vehicle for making the connections to the data.
With DataSpot, you don't have to be concerned with where the data is coming from, because it can connect with any ODBC data source (that's several dozen different drivers) and with ASCII text data, and it has its own drivers for a number of desktop PC formats such as Access, FoxPro, Paradox, and dBASE. You do need to consider what data is appropriate for a particular application. That's where you use the DataSpot Publisher to set data properties. With properties, you can selectively add and remove tables or fields from the Web View. You can also give tables and fields more user-friendly names, create concatenated fields, and set up field formatting. DataSpot does not, however, support calculated fields.
DataSpot lets you select which fields to display and their relative positioning. If you want to enhance the appearance of fields or their labels, you'll need to use HTML code. It's difficult to be creative with the display pages. DataSpot doesn't provide a WYSIWYG editor, nor does it give you convenient hooks into ActiveX controls, Java applets, or other whiz-bang page enhancements. Although capable of displaying images, DataSpot is clearly a text-oriented product, which some people will find constricting. That said, I'm sure that most of the corporate information you'd care to tap with DataSpot is plain data -- text and numbers -- which mostly needs to be formatted, not decorated, for effective presentation.
When the layout work is done, a parameters file is created that will be compiled with the Web View when the application is created. As you can hopefully tell, none of the data preparation is difficult for anybody with some background in data management. So much of the nitty-gritty, especially the details of dealing with ODBC and other database drivers, is done by DataSpot that you can afford to concentrate on proper design and configuration -- something of a luxury these busy days.
DataSpot comes with one thin manual and no online tutorial material. You might be inclined to think it's kind of cheap for a fairly expensive program, but in the end you'll understand -- DataSpot doesn't need a big manual.
To set this up, you can selectively apply what DataSpot calls "associative modules" to the Web View. These are various algorithms that scan the data and produce word and phrase associations that are later turned into hypertext URLs and search indexes. The list of these modules says a great deal about the sophistication of the DataSpot approach:
The final steps in preparing a DataSpot application are to generate the Web View and the results display template file. Depending on the number and volume of data sources, combined with the number of associative algorithms you've selected to run, Web View generation can be a lengthy process. DTL isn't kidding when it says this program needs a Pentium or Pentium Pro processor and lots of memory. The same can be said for the Web server environment that will run the applications. Response times for most queries are acceptable, but I don't think that the current version of DataSpot will develop a reputation for speed.
If you need to tweak the output display (and with a non-WYSIWIG system, who wouldn't?), you'll have to regenerate the display template file. This fix-and-generate cycle used to be common practice, but in the days of "two-way editing" and instant compiling that you find in client/server development environments, DataSpot makes you go through a lot of steps. It's the only part of the program that can be said to require some expertise and intensive labor.
Once the Web View has been generated, making the DataSpot application available to users takes nothing more than inserting an URL into an HTML document -- for example, <A HREF= "sample/wines/wines_wa.adb?">. The DataSpot server intercepts these calls and processes the Web View file. The server software has a simple Manager Window, which you can use to set operating parameters. (See Figure 3.) I also found the operating statistics it provides to be useful in determining the effectiveness of an application. You can watch query processes at work and see how many results have been found.
Users take over after the DataSpot server passes HTML pages to the browser and can follow their needs and whims for information, just as they would on the grand Internet. Only this time they're looking at corporate data that may have been locked up for decades. There are thousands of potential applications -- catalogs, sales history, in-depth databases of all kinds. DataSpot strikes me as particularly well suited for Intranets, where security is less of an issue. (DataSpot doesn't provide security controls).
I know this is supposed to be a hands-on review, but with a few exceptions, you don't have a lot of hands-on with DataSpot. I remember watching the program churn -- it does take some time to generate applications even with high-powered equipment -- thinking all the while, "This is what computers are supposed to do. They should do the grunt work that turns mountains of data into molehills of information."
Think of it this way, almost nothing that DataSpot does automatically would you want to do manually: not making connections to all the databases, or setting up the search routines, or formulating the responses. Even with its layout quirks and demand for processing power, you can accomplish things with DataSpot that you wouldn't even attempt otherwise -- all without programming. Of course, you may not want or need a $9,000 (and up) solution for simple Web data publishing, but if your situation includes a lot of disparate data sources or complex information that's difficult to categorize, then few products will serve as well as DataSpot. It's a pity there isn't a less expensive version.
What did you think of this article? Send a letter to the editor.
Figure 2.
--The DataSpot Publisher modules are where you define what tables and fields are to be included and also determine which of the association algorithms will be applied to the data.
Figure 3.
--The DataSpot Server Manager gives you some very useful information about the queries in process, as well as access to operational parameters.
Nelson King is a professional software developer specializing in Internet database management applications. He has published eight books on database subjects for MIS:
Press and frequently wears a journalist's hat.
July 1997 Table of Contents | Other Contents | Article Index | Search | Site Index | Home
DBMS and Internet Systems (http://www.dbmsmag.com)
Copyright © 1997 Miller Freeman, Inc. ALL RIGHTS RESERVED
Redistribution without permission is prohibited.