I heard about the new release from Blendernation before it was Slashdotted, and I could already see the countless trolling of "Blender's GUI isn't like my other 3D app" or "Maya's had that feature for years". Yadda yadda.
Feature-wise, Blender has many advantages over the other apps. Sometimes you can be a master at being a jack-of-all-trades (ironically), regardless if you are the best at any one thing. Blender is a production pipeline from top to bottom, including compositing that is integrated with the rest of the rendering pipeline and video editing. Blender also has decent built-in support for using Python to access its internals. Python alone is a great language, capable of serious programming from filesystems (e.g. using FUSE) to net protocols (e.g. Twisted), and the possibilities with Blender are endless. Sorry, but Maya still doesn't have comparable scripting power like this at its disposal. The point is not to say that Maya is bad, but rather that Blender has a lot of things going for it that makes it worthwhile in its own right.
Aside from feature comparisons, Blender's advantage is that it's open source. Maybe Maya will stay popular, maybe it won't, or maybe it'll get acquired by another company that might ruin it as a product. An interesting thing to notice is that community enthusiasm is different between proprietary and open source apps. Maya users tend to be enthusiastic about using Maya because of its feature set or becuase it's considered "professional". But I've seen some users turn on Maya once they started using ZBrush, probably because it gives them an easier way to make bragging-rights-quality organic models. Blender users like Blender because it is different than professional apps in its philosophy, and because they see that they have a role in Blender's future potential. Sometimes even the pure novelty of such a great project being open source is enough to keep users interested.
I used to think that way myself. It's not that it's not a true statement, it's just that the there's a greater truth to be observed concerning taking our claims in context of why we made such claims. "Knowing enough" is simply to say, "well, if I'm ever wrong, just add more knowledge to my brain until I couldn't have been wrong." It might be a compelling argument if a fact is known a priori, but it isn't much of a discovery. In practice, the more the understanding of a phenomenon approaches requiring holistic understanding (i.e. holistic in terms of how more of the universe as a whole must be taken into account), the less we can manage understanding it. The only way we can "know enough" is if we are extremely modest about what we are trying to predict. We must take an inherently undivided world and decide how to partition it for the sake of manageable thought. For example, it's no surprise that the experiment must limit the variables to only one or two things, such as temperature. Despite the idiocy of many Intelligent Design claims, many in our culture have a hard time resisting the idea that attributes are "for" something (i.e. fur is for the sake of keeping warm, etc.). Predicting evolution gets much harder when you have to take all the things that can happen in this universe into account that can cause variations that propagate, without seeing things in terms of teleology.
This isn't a mere matter of the speed of your connection to your ISP. Don't say "the internet is already tiered" because you can choose to have dial-up vs. DSL or cable modem. That's missing the point. Your overall connection speed is your overall connection speed, which affects the speed of all the pages you browse. There isn't some imposed mechanism (i.e. tiered services) that discriminates the speed before it gets to you due to preferential treatment to higher-paying web sites. Quality of Service, etc. itself doesn't already create a tiered internet. The issue is, will the tiered internet discourage openness of the internet because it favors the wealthy? "Neutrality" of the internet is good because it allows the internet to remain open. Open communication, in turn, is essential to a free and just society. What exactly do you think will happen when only the rich and powerful have a voice on the net, while the marginal whistle-blowers (who are inevitably poorer and less powerful, which is why they are in the margins in the first place)can't get the word out because they've been choked from the net?
Capitalism is the best thing we have right now, but I don't think people should be sheep. I'm tired of hearing how Capitalism is great because it encourages competition that results in better goods and services, when in reality it has nothing to do with competition. Corporations don't go out of their way to encourage competition out of the kindness of their heart, etc. It's all about the elimination of competition by the most economically powerful, perhaps at any cost (i.e. our freedom). Take the lie and make it true. Do you want competition in the market so that we aren't forced into monopolies? Then level the playing field by discouraging a tiered internet so that you don't have to be rich to be heard on the net. To me, it isn't even a question of whether or not the internet should be tiered.
So now the question is, do we think the government can do this competently, or do we have faith that neutrality can be preserved without intervention?
Most of the reasons stated by ATI and Nvidia as to why they keep drivers proprietary are for P.R. reasons, designed to be less offensive than simply saying, "we want a profit and we don't think open sourcing is worth it". Talented developers are talented developers, regardless if they work with proprietary or open source software.
The problem with proprietary drivers is that they are distributed in binary form, which turns a political problem into a technological one. If the drivers were open source, there wouldn't be a technological side to the problem (for the most part). It doesn't matter what variation of kernel you use, just compile the driver for that kernel just like you would with any other driver that's already open source.
The openness of the kernel is one of the reasons why a standard driver API hasn't been that important yet: source code is more malleable than binary. The cross-platform nature of many things concerning Linux is a testament to that.
So, in all honesty, the problem is simply 100% political, not technological. The challenge is to convince vendors that open sourcing their drivers is the better way to sell their hardware.
Re:It's funny that you should say that.
on
Pro C#
·
· Score: 1
I mostly agree. But sometimes documentation on the web leaves out a few steps and fails to show why you should prefer method x over method y when doing task z, even though you've seen it in at least a dozen code-snippet examples on the web. For a broad, coherent view, often a book is better (web tutorials usually aren't long enough or don't go into enough depth), though for specific tasks a forum is better. Besides, often I'll go to Kinko's and print out the PDF version of an API just because I like to have something on paper in front of me. I hate reading a 200 page PDF. That alone probably makes computer books worthwhile for me. Your mileage may vary.
AJAX makes pages more dynamic and responsive by alleviating page refreshes, but it still doesn't solve a fundamental problem with web apps versus traditional binary apps. The problem is, events can be initiated by the browser (ie. client) and responses are then created by the web server, but it doesn't happen the other way around: even if an event is created by the web server, the web browser isn't listening. Nothing happens unless an event is triggered from the browser (i.e. user pressed submit button, clicked a URL, or a timed script generates an event). A traditional binary asynchronously-networked app can do this. AJAX is a start in the right direction, but web apps aren't there yet. What I would like to see is true asynchronous two-way communication where either the server or browser can initiate events. The only project I've heard of that might be trying this is LivePage with Nevow (a project based on the Python Twisted web server), but maybe I'm wrong.
Most people will use a separate 3D package to output RIB, and therefore never deal with Renderman directly. However, the Renderman specification allows for language bindings, such as the official bindings for C, that you would use to have greater control over the scene. I personally am a Python user, and prefer to use the CGKit language bindings.
Slashdot Mirror
I heard about the new release from Blendernation before it was Slashdotted, and I could already see the countless trolling of "Blender's GUI isn't like my other 3D app" or "Maya's had that feature for years". Yadda yadda.
Feature-wise, Blender has many advantages over the other apps. Sometimes you can be a master at being a jack-of-all-trades (ironically), regardless if you are the best at any one thing. Blender is a production pipeline from top to bottom, including compositing that is integrated with the rest of the rendering pipeline and video editing. Blender also has decent built-in support for using Python to access its internals. Python alone is a great language, capable of serious programming from filesystems (e.g. using FUSE) to net protocols (e.g. Twisted), and the possibilities with Blender are endless. Sorry, but Maya still doesn't have comparable scripting power like this at its disposal. The point is not to say that Maya is bad, but rather that Blender has a lot of things going for it that makes it worthwhile in its own right.
Aside from feature comparisons, Blender's advantage is that it's open source. Maybe Maya will stay popular, maybe it won't, or maybe it'll get acquired by another company that might ruin it as a product. An interesting thing to notice is that community enthusiasm is different between proprietary and open source apps. Maya users tend to be enthusiastic about using Maya because of its feature set or becuase it's considered "professional". But I've seen some users turn on Maya once they started using ZBrush, probably because it gives them an easier way to make bragging-rights-quality organic models. Blender users like Blender because it is different than professional apps in its philosophy, and because they see that they have a role in Blender's future potential. Sometimes even the pure novelty of such a great project being open source is enough to keep users interested.
All this taken together, Blender is quite unique.
I used to think that way myself. It's not that it's not a true statement, it's just that the there's a greater truth to be observed concerning taking our claims in context of why we made such claims. "Knowing enough" is simply to say, "well, if I'm ever wrong, just add more knowledge to my brain until I couldn't have been wrong." It might be a compelling argument if a fact is known a priori, but it isn't much of a discovery. In practice, the more the understanding of a phenomenon approaches requiring holistic understanding (i.e. holistic in terms of how more of the universe as a whole must be taken into account), the less we can manage understanding it. The only way we can "know enough" is if we are extremely modest about what we are trying to predict. We must take an inherently undivided world and decide how to partition it for the sake of manageable thought. For example, it's no surprise that the experiment must limit the variables to only one or two things, such as temperature. Despite the idiocy of many Intelligent Design claims, many in our culture have a hard time resisting the idea that attributes are "for" something (i.e. fur is for the sake of keeping warm, etc.). Predicting evolution gets much harder when you have to take all the things that can happen in this universe into account that can cause variations that propagate, without seeing things in terms of teleology.
This isn't a mere matter of the speed of your connection to your ISP. Don't say "the internet is already tiered" because you can choose to have dial-up vs. DSL or cable modem. That's missing the point. Your overall connection speed is your overall connection speed, which affects the speed of all the pages you browse. There isn't some imposed mechanism (i.e. tiered services) that discriminates the speed before it gets to you due to preferential treatment to higher-paying web sites. Quality of Service, etc. itself doesn't already create a tiered internet. The issue is, will the tiered internet discourage openness of the internet because it favors the wealthy? "Neutrality" of the internet is good because it allows the internet to remain open. Open communication, in turn, is essential to a free and just society. What exactly do you think will happen when only the rich and powerful have a voice on the net, while the marginal whistle-blowers (who are inevitably poorer and less powerful, which is why they are in the margins in the first place)can't get the word out because they've been choked from the net?
Capitalism is the best thing we have right now, but I don't think people should be sheep. I'm tired of hearing how Capitalism is great because it encourages competition that results in better goods and services, when in reality it has nothing to do with competition. Corporations don't go out of their way to encourage competition out of the kindness of their heart, etc. It's all about the elimination of competition by the most economically powerful, perhaps at any cost (i.e. our freedom). Take the lie and make it true. Do you want competition in the market so that we aren't forced into monopolies? Then level the playing field by discouraging a tiered internet so that you don't have to be rich to be heard on the net. To me, it isn't even a question of whether or not the internet should be tiered.
So now the question is, do we think the government can do this competently, or do we have faith that neutrality can be preserved without intervention?
Most of the reasons stated by ATI and Nvidia as to why they keep drivers proprietary are for P.R. reasons, designed to be less offensive than simply saying, "we want a profit and we don't think open sourcing is worth it". Talented developers are talented developers, regardless if they work with proprietary or open source software.
The problem with proprietary drivers is that they are distributed in binary form, which turns a political problem into a technological one. If the drivers were open source, there wouldn't be a technological side to the problem (for the most part). It doesn't matter what variation of kernel you use, just compile the driver for that kernel just like you would with any other driver that's already open source.
The openness of the kernel is one of the reasons why a standard driver API hasn't been that important yet: source code is more malleable than binary. The cross-platform nature of many things concerning Linux is a testament to that.
So, in all honesty, the problem is simply 100% political, not technological. The challenge is to convince vendors that open sourcing their drivers is the better way to sell their hardware.
I mostly agree. But sometimes documentation on the web leaves out a few steps and fails to show why you should prefer method x over method y when doing task z, even though you've seen it in at least a dozen code-snippet examples on the web. For a broad, coherent view, often a book is better (web tutorials usually aren't long enough or don't go into enough depth), though for specific tasks a forum is better. Besides, often I'll go to Kinko's and print out the PDF version of an API just because I like to have something on paper in front of me. I hate reading a 200 page PDF. That alone probably makes computer books worthwhile for me. Your mileage may vary.
AJAX makes pages more dynamic and responsive by alleviating page refreshes, but it still doesn't solve a fundamental problem with web apps versus traditional binary apps. The problem is, events can be initiated by the browser (ie. client) and responses are then created by the web server, but it doesn't happen the other way around: even if an event is created by the web server, the web browser isn't listening. Nothing happens unless an event is triggered from the browser (i.e. user pressed submit button, clicked a URL, or a timed script generates an event). A traditional binary asynchronously-networked app can do this. AJAX is a start in the right direction, but web apps aren't there yet. What I would like to see is true asynchronous two-way communication where either the server or browser can initiate events. The only project I've heard of that might be trying this is LivePage with Nevow (a project based on the Python Twisted web server), but maybe I'm wrong.
Most people will use a separate 3D package to output RIB, and therefore never deal with Renderman directly. However, the Renderman specification allows for language bindings, such as the official bindings for C, that you would use to have greater control over the scene. I personally am a Python user, and prefer to use the CGKit language bindings.