Web Grid

By: Chris Sakamoto


My first inital use of the Web Grid was not a pleasant experience. I did not truly understand the reasons behind the Web Grid nor did I understand the purpose of what the Web Grid was to accomplish. Therefore my initial grid tended to be a little less focused on the constructs used, but rather it focused on how to interact and use the web gird. It also forced me to read the paper on Web Grid so that I could understand some of the terminology and reasons behind Web Grid.

The web I created in exploration was on sports...and my context was Competitive Sports. The constructs varied from Continuous/Interupted sports to outdoor/indoor sports. I rated 12 sports on 7 different levels. I discovered how to "break matches" between what was deemed to be similar by creating levels (7 of them) and if a level was deemed to be similar it could be broken by introducing a new sport. From the FOCUS analysis I found that Raquetball and Squash were identical in the levels I used, and had a correlation with Tennis. Cricket and Football were related as was speed and figure skating. Highly related levels were Continuous/Interupted with Fast Paced/Slow Paced and both of these are somewhat related to Indoor/Outdoor sports. The first two make sense, however I guess in the sports I rated I suppose that they could have been related to the indoor/outdoor situation. Using the PrnCom analysis, I got two clusters of levels...possbily typified by Raquet/Non-Raquet sport and Interrupted/Continuous sports.


In this section, I had a little better grasp on what to accomplish using Web Grid. I focused most of my time upon tring to develop useful contstructs in which were related to the Web Grid to be compared. I developed the contructs by reviewing the differences between each of the topics given in the course trying to relate them to each other in some method. For example, i felt that using a Software/Hardware level would be appropriate as I know that some of the topics focused on hardware and other on software. After rating the items on 5 levels I ran my new grid through the comparison and discovered that one of the levels for rating was exactly the same level used by the original grid. This was of application versus development tool, although the rating of the topics did not agree totally on this level. Using the terminology of a 4-way classification contstruct, these two distinctions were similar enough to form a consensus, that is same terminolgy used in the same way. The highest corresponding level was between my title of software/hardware and the original grids only act as programmed/semi-autonomous. Therefore this can be termed a correspondance, or the use of different terminology for the same distinction. Other correpondances occur between Techlogically limited/Kowledge limited of my grid and Conventional System/Intelligent System of the original grid. Production/Tool of my grid and Conventional communicatoin/Novel Communication of the original grid can be said to be in contrast, that is, having different terminology for different distinctions. As far as rating on the constructs go, the similar topics between the original grid and the one created by myself show a high correspondance between Electric Publishing, and some correspondance between Object Oriented Systems, Visual programming and Multimedia/Hypermedia. Link to grid.

Presentation Topic

Actually this was the most difficult of the 3 Web Grids to create. Creating constructs were difficult in that most constructs created either didn't apply to all the topics assigned, or were so black and white that there weren't any levels to assign between the two. However, this could also be helpful in determing differing contructs in that if comparisons are made, it could be helpful to show major differences between two sets of constructs (and obvious similarities).

This web grid focuses on 7 different Visual Progarmming Languages and rated them upon the following levels: Requires text/Visual Widgets only, Own Environmnt/Needs Envrironment, Basic/C++, Generates Code/Visual, RuleBased/Simulation. Note that Basic/C++ will only account for two of the languages involved, however, this was necessary to break a similarity difference between these two languages...although as far as Visual Languages go, this was not necessary. Since Visual C++ and Visual Basic are not true programming languages, the could have been lumped together into one category, however I was curious as to the effects of rating only 2 of the topics and leaving the remaining topics unrated.

The FOCUS analysis shows that there isn't too much correlation between the levels used to rate the topics. There weren't too many levels used either, as I had difficulty in trying to find more then those given to distinguish the differences between the Visual Programming Languages. The closest correlation between the levels is between Generates Code/Purely Visual and Requires Test/Visual Widgets only. This makes sense, as they are somewhat similar. The FOCUS analysis also shows that Visual Basic and Visual C++ are related to none of the other languages and somewhat related to themselves, as to be expected. VPlus and Prograph are shown to be exactly alike and a match breaker should be run to break apart these two languages.

The PrinCom analysis was extremely sparse due to the small degree of levels used, but shows an very interesting trend. All the levels are very bipolar, and each topic is itself rated at one end of a spectrum or another. It shows that languages that have Visual Widgets also are simulation based and purely visual (which is expected as well from researching Visual Languages). However some can be rule-based and some can generate textual code, but these are a minority. Visual C++ and basic are shown to be completely out of the realm of Visual Programming and that is to be expected.


Overall, I found the use of Web Grid to be an interesting experience. It initally was not seen as useful, but with further playing around with the grid, I found the usefulness if comparing analysis or information with others is necessary. What would be important would be the ability to filter the constructs to allow the information from several people to coincide and allow a better anaylsis of the information rather then the constructs used.

I did find a bug that was somewhat irritating. In the Web Grid, if a PrinCom or FOCUS analysis was clicked upon, it continued to print that initial analysis unless the memory cache within Netscape was deleted. This is probably a fault from Netscape, so there might be nothing that can be done on the server end, however hopefully in the future this can be remedied.

Back to my Home Page