hi zarkow,
yeah i have been following xnurbs too. it looked interesting at first. then the dev started acting poorly on the forum and turned a lot of people off. like frenchy mentions, the price is high, just to do a few things that rhino can't. xnurbs homepage was always a red flag for me. he constantly claims it's the best nurbs kernel around, when it's obviously not a nurbs kernel. it just does some things a nurbs kernel can't do yet. if everyone wasn't so money hungry, xnrubs would have just been a patch to an open source nurbs kernel. it looks like a useful product though. i think the price makes your dream of moi + xnurbs impossible. he is charging so much for so little. not sure how he could ever sell it to moi customers.
i think the easiest thing for me would be to stay with rhino, since i'm familiar with it, and add either moi or plasticity to get decent filleting. the advanced surface stuff isn't necessarily something i would have to have. i spent a few months trying to make a traditional fixed pitch aircraft propeller (Sopwith Camel) in rhino. i constantly ran into bugs/problems, until i finally got a workflow that gave reasonable results. in that case, early on, more advanced surfacing methods might have helped. hard to say for sure though.
attached, are pics of the traditional fixed pitch aircraft propeller that i created in rhino 5. getting the transitions between the blades and hub to match how they are actually made was the challenge. even after months of banging on rhino, the method i eventually found is still limited by fillet size and doesn't always work. it depends on a lot of things. it's just so buggy and frustrating. i should mention the method i eventually found was based on rhino not letting me do anything else. so i just kept running into problem after problem, until i found at least one way to progress. so it wasn't a method i created by some logical reason.
in your video, where you were having problems making certain surfaces, i run into that situation sometimes too. in most of your video, you were doing things i didn't even know about. i am always working from fixed points that i import. so i never knew about just moving lines around and having surfaces follow them. that looks like an enjoyable way to work. i just don't have that freedom, unfortunately. i also have to be able to export step files that will import into netgen. that is where moi had problems. rhino doesn't have that problem, for some reason. i haven't tested plasticity yet. just been following it's dev.
in the part of your video, where you couldn't get extended surfaces. i have had some success with that, but it's for short distances. when the distance grows, then things go crazy. it's the same with splines too. it's nice that plasticity doesn't seem to have that problem.
the new moi acis fillet is an exciting upgrade for me. so i'm curious to see how that goes. rhino always fails or limits the size of the fillets i can do. i eventually found fillet surface works a little better than fillet edge. but they are both problematic.
i was reading the history of rhino, an interview with michael, and a video interview of bob mcneel. bob flat out said, with no qualms, that he ignores user requests for years and only does stuff that will ensure he's making money and keeping the lifestyle he is accustom to. so fits perfectly with what i've seen out of rhino. high prices, very slow dev, and ignored users. just a few weeks ago, they fixed a bug reported 10 years ago. michael's interview was a lot better. he says he likes listening to users and finding solutions to their underlying issues. which you can clearly see from moi's dev history. michael also said something interesting that i see all the time. users always ask for mutually exclusive stuff. some are adamant they need some feature while other users are adamant they don't want it. so it's not possible to listen to every user all of the time. thus, trying to fix the underlying issues is the best path. makes total sense. michael also stated his goal was to provide easy to use, affordable software, for an under served community. you can clearly see that he's made good on that as well.
i think, right now, the best thing for me to do would be make the basics in rhino, then import that to moi or plasticity, and do the filleting. if using moi, i would have to export back to rhino (since i need a step file that works with netgen). plasticity may not have that problem. so i could go rhino, to plasticity, to step, then into netgen. but all the extra software because rhino's kernel sucks. they've had forever to develop it. it truly seems hopeless at this point. i have to laugh, every time i read on their forum, when they say well kernel dev is hard, it takes time, be patient. christ, i'm retired now, have been for 15 years. plus all the years of working. they require so much patience that you have to wait until your dead before it works right. this is where plasticity really excels the most. he did all that in a year.
the kernel history is interesting too. acis and parasolid both came from the same group of british devs. all the nurbs kernels came from the same group of, i believe, canadian devs. all the kernel dev companies have been bought and sold many times. so they all took different dev paths. but there are basically two types created by two different teams. it's clear that the nurbs kernels are inadequate and at the pace they make improvements, it's not good. so the real battle is between acis and parasolid. they have progressed well too. i'm not sure which is best. the plasticity dev has a video where he says parasolid is superior. i think either one is better than any of the nurbs kernels though. the plasticity dev had the right idea. just find out which is the best kernel and use it. don't do what rhino did. by an old kernel and claim to be improving it, but not mention it's going to take 100 years. i mean really, wtf. michael has an interesting approach. combining a nurbs kernel with acis. he's also using what appears to be the best of the nurbs kernels. integrityware, fairly recently, added a C# version of the kernel. that could be something that helps improve things too. not really sure.
so all that rant is about your last big point in your video. the killer features you want in moi are all because of the crappy nurbs kernel. plasticity has what you want because of the superior parasolid kernel. however, you don't like some of the things it's missing. those are things the nurbs kernel can do. so michael's dual kernel approach might give you everything you want, one can hope. interestingly, some cad programs use the parasolid and acis kernels, in the same product.
well, anything is better than waiting on rhino to wake up. bob's too busy enjoying his grand lifestyle, apparently. at least he's modest, and trying to hide it, lol. nah, let's just rub it in. in the end of his interview, his advise to students was to follow the money. what a guy. thank god the entire world isn't like him.
anthony
|
.png)
