3 ms·

Cool, thanks for sharing that. I had no idea Andy Garvin was on hn, but makes sense :)! I had a vague memory that Super Mario 64 also used lisp and when I goog

by rgbrgb 5y ago

Cool, thanks for sharing that. I had no idea Andy Garvin was on hn, but makes sense :)! I had a vague memory that Super Mario 64 also used lisp and when I googled, I found him again [0]. You've likely seen this, but I recently watched and loved this extended interview with him about making Crash Bandicoot [1]. Really made me want to look more into making video games. [0]: https://news.ycombinator.com/item?id=2474833 https://news.ycombinator.com/item?id=2474833 [1]: https://www.youtube.com/watch?v=pSHj5UKSylk https://www.youtube.com/watch?v=pSHj5UKSylk

Jach 5y ago
Yeah it's rather amazing what Lisp applications in the 90s contributed to. https://en.wikipedia.org/wiki/N-World https://en.wikipedia.org/wiki/N-World is the software mentioned in the comment, the modern equivalent being Blender. The Credits section of that page has some great classics like SM64, FF7, Banjo-Kazooie... It also had great tutorials and documentation: http://www.aaronjamesrogers.com/misc/nworld/N-World-Intro.html http://www.aaronjamesrogers.com/misc/nworld/N-World-Intro.ht... It's a shame that its successor Mirai https://en.wikipedia.org/wiki/Mirai_(software) https://en.wikipedia.org/wiki/Mirai_(software) just kind of vanished. Its main claim to fame is Gollum's face. Chronologically it's not that different from Blender, which also imploded around the same time in the early 2000s, but Blender was able to go open source and survive and have another 20 years of development to get where it's at today. I don't know that the world would look all that different if Mirai had been able to take a similar path, but at least there'd be a cool modern 3D modeling tool in Lisp.
- deterministic 5y ago
  I worked on Banjo Kazooie and the claim that N-World was used for it is incorrect. We used Maya for all modelling. They did a demo to the BK team once and it went really badly for them. What they demoed to us was so far from being competitive with Maya that it was almost laughable.
  - Jach 5y ago
    Thanks for the correction! I guess another instance of Rare just doing it better than Nintendo ;) I've been trying to track down any source for the original rumor without success, the random IP seems to have made it up. Did you ever get to see much of their later Mirai tool? Though I would also believe Maya was quite far ahead at that time too.
    - deterministic 5y ago
      No the demo went so badly for them that we never invited them back. I kinda felt sorry for them. They had no clue how uncompetitive they were compared with the tools we were already using. One of the two demo guys kept going on about how better it was because it was using Lisp. While also acknowledging that it was missing pretty much all the features we needed. The “it’s easy to implement yourself because Lisp!” argument was laughable given that Maya already had the features we needed and used every day.
      - Jach 5y ago
        > They had no clue how uncompetitive they were compared with ... Sadly not that uncommon in Lisp advocates for a number of different things even today... though I think that's been slowly changing as more Lispers either become aware of what's out there, or come with experience of what's out there and bring that knowledge with them to Lisp. The reverse direction seems to happen at least as slowly -- I've commented before that a JRebel setup is very close to bringing the still largely unique interactive development mode of Lisp over to the Java world, but it's also in important ways not fully there and may never be. Meanwhile many other languages don't even care about that development style at all. > The “it’s easy to implement yourself because Lisp!” argument was laughable I guess this would be kinda analogous to when Blender first added Python support, if they told everyone to just do expected things themselves. It may be true that such-and-such task is relatively easy to implement (if indeed it's possible to do at all without changing something about the application itself) but it's also just a terrible thing to focus sales on when the value seems like it'd mostly be measured by how useful the tool is to non-programmers. I hadn't really looked much at the "Wide Open World" doc on the archive since it's not mentioned in the other docs except vaguely in like one spot I had to search for, but if a customer had heavy scripting needs to accomplish things not already provided or easily done with the graphical tools, now the customer's not only being sold on a graphics tool but also on learning to program in Lisp using their bundled xemacs. And as you note it's even worse when that work is already done in another tool that can't be integrated with. Though I see they provided an FFI so there could be a possibility of integrating C code the customer or others wrote. If a customer was expecting to have to do a bunch of new custom stuff regardless of tool maybe it'd have made more sense... But now at least I have an additional insight into why they weren't very successful. I'm appreciating anew a lot of things the last products I worked on got right in terms of providing a good enough and frequently improving out of the box set of components and workflows while paying some attention to competitors to see what key things were lacking on either side, while streamlining development and deployment of custom components, and also providing a platform for customers to share their own custom work with each other either openly or by putting a price on it.

by rgbrgb 5y ago

Jach 5y ago

Yeah it's rather amazing what Lisp applications in the 90s contributed to. https://en.wikipedia.org/wiki/N-World https://en.wikipedia.org/wiki/N-World is the software mentioned in the comment, the modern equivalent being Blender. The Credits section of that page has some great classics like SM64, FF7, Banjo-Kazooie... It also had great tutorials and documentation: http://www.aaronjamesrogers.com/misc/nworld/N-World-Intro.html http://www.aaronjamesrogers.com/misc/nworld/N-World-Intro.ht... It's a shame that its successor Mirai https://en.wikipedia.org/wiki/Mirai_(software) https://en.wikipedia.org/wiki/Mirai_(software) just kind of vanished. Its main claim to fame is Gollum's face. Chronologically it's not that different from Blender, which also imploded around the same time in the early 2000s, but Blender was able to go open source and survive and have another 20 years of development to get where it's at today. I don't know that the world would look all that different if Mirai had been able to take a similar path, but at least there'd be a cool modern 3D modeling tool in Lisp.

deterministic 5y ago
I worked on Banjo Kazooie and the claim that N-World was used for it is incorrect. We used Maya for all modelling. They did a demo to the BK team once and it went really badly for them. What they demoed to us was so far from being competitive with Maya that it was almost laughable.
- Jach 5y ago
  Thanks for the correction! I guess another instance of Rare just doing it better than Nintendo ;) I've been trying to track down any source for the original rumor without success, the random IP seems to have made it up. Did you ever get to see much of their later Mirai tool? Though I would also believe Maya was quite far ahead at that time too.
  - deterministic 5y ago
    No the demo went so badly for them that we never invited them back. I kinda felt sorry for them. They had no clue how uncompetitive they were compared with the tools we were already using. One of the two demo guys kept going on about how better it was because it was using Lisp. While also acknowledging that it was missing pretty much all the features we needed. The “it’s easy to implement yourself because Lisp!” argument was laughable given that Maya already had the features we needed and used every day.
    - Jach 5y ago
      > They had no clue how uncompetitive they were compared with ... Sadly not that uncommon in Lisp advocates for a number of different things even today... though I think that's been slowly changing as more Lispers either become aware of what's out there, or come with experience of what's out there and bring that knowledge with them to Lisp. The reverse direction seems to happen at least as slowly -- I've commented before that a JRebel setup is very close to bringing the still largely unique interactive development mode of Lisp over to the Java world, but it's also in important ways not fully there and may never be. Meanwhile many other languages don't even care about that development style at all. > The “it’s easy to implement yourself because Lisp!” argument was laughable I guess this would be kinda analogous to when Blender first added Python support, if they told everyone to just do expected things themselves. It may be true that such-and-such task is relatively easy to implement (if indeed it's possible to do at all without changing something about the application itself) but it's also just a terrible thing to focus sales on when the value seems like it'd mostly be measured by how useful the tool is to non-programmers. I hadn't really looked much at the "Wide Open World" doc on the archive since it's not mentioned in the other docs except vaguely in like one spot I had to search for, but if a customer had heavy scripting needs to accomplish things not already provided or easily done with the graphical tools, now the customer's not only being sold on a graphics tool but also on learning to program in Lisp using their bundled xemacs. And as you note it's even worse when that work is already done in another tool that can't be integrated with. Though I see they provided an FFI so there could be a possibility of integrating C code the customer or others wrote. If a customer was expecting to have to do a bunch of new custom stuff regardless of tool maybe it'd have made more sense... But now at least I have an additional insight into why they weren't very successful. I'm appreciating anew a lot of things the last products I worked on got right in terms of providing a good enough and frequently improving out of the box set of components and workflows while paying some attention to competitors to see what key things were lacking on either side, while streamlining development and deployment of custom components, and also providing a platform for customers to share their own custom work with each other either openly or by putting a price on it.