mosko@leconte.math.ucla.edu (Bradley Moskowitz) (02/06/91)
It seems to me at first glance that AVS from Stardent is very similar to apE in its approach. Is there any connection between them, e.g. one based on the other in concept, or are they two very different systems ? Thank you for any comments in this regard. Brad M.
ktayama@bonnie.ics.uci.edu (Katsumi Tayama) (02/07/91)
In article <1044@kaos.MATH.UCLA.EDU> mosko@leconte.math.ucla.edu (Bradley Moskowitz) writes: >It seems to me at first glance that AVS from Stardent is very similar to apE >in its approach. Is there any connection between them, e.g. one based on the >other in concept, or are they two very different systems ? Thank you for any >comments in this regard. > According to: Dyer, D. Scott A dataflow toolkit for visualization (apE) IEEE Computer Graphics and Applications 10: 60-9 Jl '90 the two systems are developed completely independently. See also: Upson, Craig Faulhaber, Thomas Kamins, David The Application Visualization System: a computational environment for scientific visualization IEEE Computer Graphics and Applications 9: 30-42 Jl '89 Kats ktayama@ics.uci.edu
wave@media-lab.MEDIA.MIT.EDU (Michael B. Johnson) (02/07/91)
In article <1044@kaos.MATH.UCLA.EDU> mosko@leconte.math.ucla.edu (Bradley Moskowitz) writes: >>It seems to me at first glance that AVS from Stardent is very similar to apE >>in its approach. Is there any connection between them, e.g. one based on the >>other in concept, or are they two very different systems ? Thank you for any >>comments in this regard. >> >>Brad M. Having reasonably extensive experience with AVS (I've written quite a number of my own modules over the last year or so) and just a bit of experience with apE (just got apE 2.0 a few weeks ago, went through the tutorials on an SGI, listened to a coworker complain about his lack of success getting it to run on an HP-9000/835, had little luck compiling it on my Stardent Titan), I would say that AVS comes across as a much more polished and more responsive system than apE. As for connection between the two systems, I don't think there is one. The notion of data flow and graphical languages applied to scientific computation has been around for quite some time in the research community, and both apE and AVS are several year old, admirable efforts to get it out to a user community. I like AVS a lot, but am disappointed with several things about it, none of which (to my knowledge, at least) have been any more successfully addressed by apE. Being able to take advantage of locally resident data with shared memory segments, encapsulation of networks (AVS) or pipelines (apE) to promote functional abstraction, or even a "network compiler"... All of these need to be provided at some point. apE does have the definite advantage that, although it is not in the public domain, the source code is readily available for a very small fee, it runs (to a greater or lesser degree) on many different platforms, and the folks at the Ohio Supercomputer Center seem to be genuinely behind their hh line is becoming the assumed platform for high-end visualization, it is very nice that apE runs on it (while AVS does not - an obvious marketing decision on Stardent's part for which I can't really fault them). SGI has dropped hints about a system they will be coming out with which will run on their systems and will be a "second generation" system (where apE and AVS are "first generation", I suppose), but so far I at least haven't seen it. Anyway, this is straying, so I'll leave it at that. -- --> Michael B. Johnson --> MIT Media Lab -- Computer Graphics & Animation Group --> (617) 253-0663 -- wave@media-lab.media.mit.edu
rick@hanauma.stanford.edu (Richard Ottolini) (02/08/91)
>In article <1044@kaos.MATH.UCLA.EDU> mosko@leconte.math.ucla.edu (Bradley Moskowitz) writes: >>It seems to me at first glance that AVS from Stardent is very similar to apE >>in its approach. Is there any connection between them, e.g. one based on the >>other in concept, or are they two very different systems ? Thank you for any >>comments in this regard. Dataflow toolkits are very common in signal processing-- speech, seismic, etc. Toolkits are ubiquitis in UNIX ala text processing, databases. Where convergence might take place is when people add a function they see done well in some other's toolkit.