Who said SketchUp doesn't need to be 64 bit?
-
Thanks for the great response Thomas. I really appreciate you taking your time to allay some of my concerns.
From the sound of things, Sketchup is in good hands when it comes to enabling third party developers to push plugin development even further.
I definitely appreciate the need to avoid bloat. Everyone has a different set of core plugins they would prefer to come as standard, but we all vary in our modelling approach. Even if we were only talking about simple plugins like weld, by the time everyone chimes in with their request the list of core tools would be ridiculous.
Being able to customise Sketchup according to your focus of interest and modelling approach is something I wouldn't want to change. The flexibility is pretty much unrivalled and I can't think of another programme that I can modify as my skills improve and evolve.
The integration of plugins could be greatly improved, but given that the better ones feel like native tools, I suspect this is down to the authors rather than a problem with SU.
That being said, I do think that there is a strong demand for native non-destructive SubD, bevelling and proper UV mapping. I also think that these features are needed to keep Sketchup competitive with rival software. They may not be required when it comes to "3D for everyone" but as with the Sandbox tools, they could be disabled when not required.
Anyway, thanks again for the positive response. It's just a pity that the same message doesn't come across from product management.
-
Taking off my Trimble hat for a moment - here's a peek at my latest personal project I do on my spare time. I showed this to people at BaseCamp so this is the first preview for those of you who where not there.
http://www.youtube.com/watch?v=e1MhcTknJWI
Quads and Subdivisions
This is all done with SketchUp 2014 using the current Ruby API. Non-destructive subdivisions with soft creases that support the pseduo-quads of QuadFace Tools (if you are like me who like to make geometry in quads.) Anything else will be triangulated.
Note that this is a very early development preview - not even in alpha stage.This is the type of extension that would fit to a niche segment of SketchUp users and therefore not be fit for core product. It isn't something everyone wants or needs. This is specialisation. But as an extension it's great, as you just bolt it on and you got yourself a subdivision modeller.
I haven't profiled and optimized this thing yet - just getting it to do the right thing. But from some quick tests its not Ruby that is the main bottleneck. This means the core need to be improved, but also that extensions aren't second grade members. There are certainly lots of things in the API that can be improved to make the creation of such extensions much easier.
-
You sly dog...
-
WOW! Really great response Thomas!
I never expected to see something like that so quickly (after following the discission of the last days)Looking foreward to the release!!! I think this will be definitely the day, i will upgrade to 2014...
Really impressive work. Thank you!
-
That look very cool indeed but it's not available for us...
I'm still waiting for a release of your Bezier Surface plugin...
Do you have any time working on these things?I agree that much can be accomplished with better plugin integration but a thing like quads (and better editing without destroying UVs). Wouldn't it be better if integrated to the core program so all tools would benefit?
Otherwise there could be for example numerous 3rd party move tools that would do essentially the same thing? One with each plugin authors solution to move vertices without messing up UVs.
That could easily make SU bloated IMHO.Also there has been talks here of real curves in SU.
Do you mean that could be implemented through 3d party or would that have to be a "core" feature? -
@pixero said:
That look very cool indeed but it's not available for us...
I'm still waiting for a release of your Bezier Surface plugin...
Do you have any time working on these things?Time is a shortage, no doubt. And to be honest, Bezier Surface hasn't gotten a lot of love lately. There is a huge amount of challenges of making bezier surfaces to appear smooth and not end up with artifacts. Subdivisions on the other hand are much easier to consistently generate a smooth limit surface. You can find many white papers on these topics that mention these things. So I've been moving more toward focusing on SUbD. But quite a bit build upon the development of Bezier Surface - and an interesting property of a 3x3 patches of quads subdivided is that they normalize to a B-Spline patch - meaning an abstraction can be built on top of the subdivision tool.
-
But does this means it would be possible to export this as a subd/bezierpatch or would it be exported as polys?
-
@pixero said:
I agree that much can be accomplished with better plugin integration but a thing like quads (and better editing without destroying UVs). Wouldn't it be better if integrated to the core program so all tools would benefit?
Otherwise there could be for example numerous 3rd party move tools that would do essentially the same thing? One with each plugin authors solution to move vertices without messing up UVs.
That could easily make SU bloated IMHO.Oh, there are most certainly a number of things in the SketchUp core that can be improved. I'm not saying that things are set in stone and the core will never change. But the essence is improve the core generically, specialize with extensions.
We have a looong list on the extensibility team of improvements to the API we'd like to see. We still rely on a good connection with the community to identify new things and put everything together in the big perspective so we can yield the best bang for the bucks for each release.
Along with improvements to the API (which may require core improvements) we also want to build up a good network for developers to use. This include sharing experience and interacting with open source projects so the wheel isn't reinvented for each extension. The rate of progress will be so much faster and better if we all build upon each other.
There is admittedly a bit of a void for this atm. We have EW, API docs and a GitHub account, but lack a central hub to message this efficiently. Keep an eye out for changes to this!@pixero said:
Also there has been talks here of real curves in SU.
Do you mean that could be implemented through 3d party or would that have to be a "core" feature?That's something that would have to be a core change - SketchUp, at heart, only deal with edges and faces - everything else is an abstraction. Real curves would require some big fundamental changes. Custom entities would be a great feature though. dreams
-
@pixero said:
But does this means it would be possible to export this as a subd/bezierpatch or would it be exported as polys?
In theory, yes. One could export to a format that supported subdivisions and export only the control mesh. (Or one could manually toggle off the subdivision before exporting into a simple mesh format.)
The challenge here is that there are many different subdivision algorithms that will yield different results. I've been using Catmull-Clark - but even in that area there are different nuances in implementation, especially when you start to include soft creases.In response to this Pixar has released OpenSubDiv, a standardization of the subdivision techniques it encourages software developers to make use of so geometry can be shared without loss of data.
https://github.com/PixarAnimationStudios/OpenSubdivI only just recently came across it and I'm not sure if it'll make it into the initial release. But that would be the best solution for exchanging subdivided models and it is indeed possible.
-
Another quick question about you video demo. I see that you are doing all edits when only the control cage i visible. I assume this is a limitation to the alpha version?
-
Great stuff Thomas - very exciting
-
@pixero said:
Another quick question about you video demo. I see that you are doing all edits when only the control cage i visible. I assume this is a limitation to the alpha version?
I started a new thread for this topic: http://sketchucation.com/forums/viewtopic.php?f=323%26amp;t=57480
-
@andrews said:
Obviously these sorts of statements seem ridiculous in retrospect, but the thing is that one never really knows which assumptions will hold true and which ones won't back when the original decisions are made.
The safe bet is generally not to artificially limit things when it comes to computers. I had a 40MB full height MFM drive. I remember installing and formatting the first 400MB drive in my city. I also remember seeing the first 1GB drive come through that shop and wondering just what the heck they were going to put on that monster. Now I don't leave the house without a 32GB flash drive in my pocket and a 32GB MicroSD in my phone. Nearly every one of these devices, at its introduction, had to have some sort of workarounds for limitations in the major OS of the time, because the developers had assumed that no one would ever use that much storage.
Of course, an even more helpful approach would be to waterboard developers when they add bloat. I remember using an 8086 and 5.25" floppies to do a lot of things that now require a CD full of software and are far from instantaneous on a 2.4GHz CPU.
-
@tt_su said:
I have the impression that many people think of extensions as second grade tool sets - which is unfortunate. Maybe it's because the majority of current extensions is made by hobbyist (I have been) without the resources to provide dedicated support and full time development.
in general, that's how i feel about extensions.. before solid tools, we had bool tools. when using boolTools, it was always "ok, gotta save before using because i have a 50-50 chance of sketchup crashing after this".. when it did work, it would often take an extended period of time for the operation to complete.. there were installation issues not necessarily related to booltools (iirc, it was actually the podium installer creating a folder in a separate library but it caused boolTools&subsmooth to not load whereas other plugins were not affected).. also the toolbars wouldn't remember their state on mac.. booltools and it's toolbars were always on unless you uninstalled..
then su8 came out and its solidTools were superior.. fast and stable with none of the toolbar/install hassles.. their implementation was much more 'professional' feeling.
i'm not singling out dale here and i think he's written some of the sweeter plugins out of anybody. it's just that in the case of boolTools, they were replaced by the native version so it allows to make a comparison between 'grade a' and 'grade b' extensions.. but if i were to make an overall gripe about plugins, bool tools wouldn't be prominent on that list.
anyway- i completely understand that most plugins for sketchup are more hobbyist related.. one man operations which in most circumstances, are being done for the good of the community (i.e.- awesome)
but when i see the way plugins often work with other applications, there still is the hobby type scripts which can be requested/found at forums and such but then there's another tier of professional writers.. infrastructure, development teams, websites, support, sales, etc.. i imagine you (@tt) know what i'm saying and these are the people you'd like to attract to sketchup as a platform.
(examples-- visualArq or clayoo )i think maybe the toughest thing to deal with when/if those devs come is going to be money.. that type of development doesn't (and can't) come cheap.. i was blown away when tgi3d was (and maybe still is?) trying to sell their software for a thousand dollars.. that's just too expensive for sketchup (as in- twice the amount of the base app).. but, at the same time, i get it that they're selling it at a fair price as a 'pro' plugin writer.. i don't have a solution and i'm not trying to suggest one either (just talking is all).
dale (whaat) prices them 'right'.. $10, 20, 40 range.. i could see some plugins up in the $2-300 range, especially if they work well, but really-- anything beyond that is just not going to sell.
[EDIT]- well, i'm wrong about expensive plugins not selling.. in essence, many of the renderers are plugins and they're often above the 2-300 mark..
-
oh.. i started rambling again and forgot to ask the question i wanted to
could solidTools, exactly as we know them today, have been written by a third party or does their speed&reliability come from things which only the Sketchup developers have access to?
-
@jeff hammond said:
dale (whaat) prices them 'right'.. $10, 20, 40 range.. i could see some plugins up in the $2-300 range, especially if they work well, but really-- anything beyond that is just not going to sell.
The problem there is that they add up quickly, and a lot of them are doing things that really should have been base SU functionality. (Add Centerpoint, Selection Toys and Purge come screaming to mind.) I wouldn't want to explain to the boss why I'm wearing out the company credit card on $10-20 extensions to do fairly basic stuff.
Then there's the separate issue of frequency of use; I probably have a couple of dozen plugins downloaded that I use maybe once a month, if that.
-
@airwindsolar said:
@jeff hammond said:
dale (whaat) prices them 'right'.. $10, 20, 40 range.. i could see some plugins up in the $2-300 range, especially if they work well, but really-- anything beyond that is just not going to sell.
The problem there is that they add up quickly, and a lot of them are doing things that really should have been base SU functionality. (Add Centerpoint, Selection Toys and Purge come screaming to mind.) I wouldn't want to explain to the boss why I'm wearing out the company credit card on $10-20 extensions to do fairly basic stuff.
Then there's the separate issue of frequency of use; I probably have a couple of dozen plugins downloaded that I use maybe once a month, if that.
$10-$40 for plugins is cheap for the amount of work the authors put into them. I never asked my boss to pay for my plugins, I bought them with my own money out of respect for the plugin authors. The frequently used "basic" plugins you mention are free but I wouldn't mind paying for them either.
Can you say which $10-$20 plugins do "fairly basic stuff" because generally the ones in that price bracket are shit-hot and invaluable. They are worth double or triple what the authors sell them for, but they realise the SU community tends towards freebies and very cheap products and nothing will sell if it's over $50 or so. Other 3D communities aren't like this and people even pay for simple tutorials. Things are changing though, if you want a high quality plugin/product for SU, then you have to pay for it because people don't work for free! I realised this when I was giving away freebie models, getting tens of thousands of downloads and only receiving a few thank yous.
-
@airwindsolar said:
@jeff hammond said:
dale (whaat) prices them 'right'.. $10, 20, 40 range.. i could see some plugins up in the $2-300 range, especially if they work well, but really-- anything beyond that is just not going to sell.
The problem there is that they add up quickly, and a lot of them are doing things that really should have been base SU functionality. (Add Centerpoint, Selection Toys and Purge come screaming to mind.) I wouldn't want to explain to the boss why I'm wearing out the company credit card on $10-20 extensions to do fairly basic stuff.
Then there's the separate issue of frequency of use; I probably have a couple of dozen plugins downloaded that I use maybe once a month, if that.
that's sort of outlining the issue though..
if sketchup were to attract professional 3rd party developers (and i mean 'professional' basically in that developers whose main source of income would be that of developing/selling plugins and whose available resources for support/coding/etc would be greater than that of a one man do-it-all operation).. then people will have to pay for it.. not in the sense that they can donate or whatever.. they must pay or it wouldn't be possible within the monetary structure of today's world.if you want freebies than that's what you'll get.. if you want polished, high performance plugins then you'll have to be willing to pay for them or the developers aren't going to come to the platform.
-
Can someone please explain to me how is it that with all the improvements in 2014 SketchUp still struggles just navigating through a large model while a third party plugin like LightUp has no difficulty at all to do a high res fly through and running a screen capture software at the same time.
64 bit or not , It seems to me that even the latest SU release is still not able to take a full advantage of currently available hardware.
See the Templar Town Tour topic for a good example.
edit: forgot to include the link.
-
@tandem said:
Can someone please explain to me how is it that with all the improvements in 2014 SketchUp still struggles just navigating through a large model while a third party plugin like LightUp has no difficulty at all to do a high res fly through and running a screen capture software at the same time.
in SU you can edit the model... sothat the inference engine and automatic snapping etc. is required to permanently interact with the whole geometry, just displaying rendered raster data is a completely different thing and thus can be distributed to multiple kernels or computed by the GPU.
Norbert
Advertisement