Hi Ivan, Sorry for the late response, I guess I was taking a vacation from myself. ... More reblow
On Thu, Dec 23, 2021 at 11:33 PM Ivan V. <[email protected]> wrote: > > I also decided to change the app name to "CogServerLab", if that's ok. The > name is unique, and Google Search likes it. I just now created https://github.com/opencog/cogprotolab/ per your later emails. I could not figure out what your github ID is, so I could not give you access. What's your github ID? > I can't express my gratitude, it is the best Christmas present ever. Thank > you very much. Well, apparently, the gift is a bit late .. > Also, what license does OpenCog prefer? Almost all the repos use agplv3 > > It would be fun to load up the atomspace with one of the big datasets, e.g. > > at https://linas.org/datasets/ and prowl around in that and see what > > happens. > > It depends on the amount of RAM. The biggest issue is bitmap caching, so if > each oval takes 500 x 500 pixels in average, 4B per pixel, then it is about > 1MB per oval, meaning 1GB of free ram should be able to hold about 1000 > ovals. In practice, the shallowest oval bitmap is the biggest one, grading > down with structure depth. Maybe I could work on this, to remove caching, > then it would be more scalable because only visible ovals would take the > bitmap memory, but the rendering would be slower and blocking. I'll try to > provide both versions (cached and non-cached) switch in a separate > configuration file. Other than bitmaps, an s-expression array and its tree > structure array are required to reside in RAM all the time. It's hard for me to imagine how rendering takes more than fractions of a millisecond per bubble. > But I'm not sure that mega-sized-datasets are practical examples of using > CogServerLab. They're not mega-sized, they're "normal sized". Except for assorted toy problems and demos, datasets consisting of 1 to 100 million atoms would be "typical". They fit easily into RAM, they don't need the fancy footwork that Ben talks about. > I imagine some other uses where newbies can play with a few hundreds of atoms > while learning how AtomSpace behaves on what Scheme commands in a realtime. > It would be also great to see someone making an OpenCog introduction or > feature video using CogServerLab. Certainly, I agree CogServerLab wouldn't be > the most useful application in the world, but I guess it could be fun to play > with. Yeah, we need demos like these. We don't have enough "hello world" entry level stuff. > Otherwise, what is happening with OpenCog Wiki pages? It would be nice to > have an objective reference to CogServerLab from there. People who have logins can create wiki pages. You could too. Accounts have to be added by hand, as otherwise spammers take over. --linas -- Patrick: Are they laughing at us? Sponge Bob: No, Patrick, they are laughing next to us. -- You received this message because you are subscribed to the Google Groups "opencog" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/opencog/CAHrUA373G9E4CGqw1U%3D1BaKz5TSEZixj-MQEpct-8iZEq2-EjQ%40mail.gmail.com.
