Options

Azimuth Code Project

2

Comments

  • 51.

    Hi Jim,

    Thanks for putting all that information together. I use git in my workflow also, so I'll keep an eye on the GitHub repo. I put my wiki page together, like you suggested, see here Patrick McLaren.

    Simple fluid dynamics on a rotating planet has caught my eye :-)

    Have a great holiday!

    Comment Source:Hi Jim, Thanks for putting all that information together. I use git in my workflow also, so I'll keep an eye on the GitHub repo. I put my wiki page together, like you suggested, see here [[Patrick McLaren]]. Simple fluid dynamics on a rotating planet has caught my eye :-) Have a great holiday!
  • 52.

    Good luck, Patrick! Let us know any questions you may have! The delayed action oscillator would be great to see, and so would simple fluid dynamics on a rotating planet. I could help with the math and physics, though not with the coding.

    Comment Source:Good luck, Patrick! Let us know any questions you may have! The delayed action oscillator would be great to see, and so would simple fluid dynamics on a rotating planet. I could help with the math and physics, though not with the coding.
  • 53.
    edited June 2014

    I added this information to the page Azimuth Code Project: it's supposed to make it easier to find the online models people have created.

    Online models

    The Azimuth coding project has produced a number of online models. Allan Erskine helped to create an interactive online model of stochastic resonance which you can find here:

    Stochastic resonance example.

    This model was explained in the blog post Increasing the signal-to-noise ratio with more noise.

    Allan Erskine has also created some models of bistability. The first was for a static climate system:

    Bistable temperature equilibrium model.

    The second was for a dynamical system:

    Bistable temperature dynamics.

    Michael Knap combined the latter model with the work on stochastic resonance to get a stochastic resonance model more explicitly connected to climate physics, which is here:

    A stochastic energy balance model.

    Comment Source:I added this information to the page [[Azimuth Code Project]]: it's supposed to make it easier to find the online models people have created. ## Online models ## The Azimuth coding project has produced a number of online models. [[Allan Erskine]] helped to create an interactive online model of stochastic resonance which you can find here: &bull; <a href="http://www.adgie.f9.co.uk/azimuth/stochastic-resonance/Javascript/StochasticResonanceEuler.html">Stochastic resonance example</a>. This model was explained in the blog post <a href = "http://johncarlosbaez.wordpress.com/2012/07/30/increasing-the-signal-to-noise-ratio-with-more-noise/">Increasing the signal-to-noise ratio with more noise</a>. [[Allan Erskine]] has also created some models of bistability. The first was for a static climate system: &bull; <a href = "http://math.ucr.edu/home/baez/coalbedo/coalbedo_new.html">Bistable temperature equilibrium model</a>. The second was for a dynamical system: &bull; <a href = "http://math.ucr.edu/home/baez/coalbedo/temperature.html">Bistable temperature dynamics</a>. [[Michael Knap]] combined the latter model with the work on stochastic resonance to get a stochastic resonance model more explicitly connected to climate physics, which is here: &bull; <a href = "http://math.ucr.edu/home/baez/coalbedo/stochastic/stochastic.html">A stochastic energy balance model</a>.
  • 54.
    edited January 2013

    Patrick wrote:

    Simple fluid dynamics on a rotating planet has caught my eye :-)

    I have collected some material that is needed for this on the page Atmospheric and oceanic fluid dynamics. However, since I am not very good at programming C++ and also the equations are not that simple, I then decided to try to write a numerical solver for the Burgers equation in a case where there is also an analytical solution (see that page), one space dimension with periodical boundary conditions.

    This is far simpler, but is still a nice programming challenge, at least for me. Unfortunately I haven't had any time for this project during the past months, but I hope I can pick it up again during the next months.

    Comment Source:Patrick wrote: <blockquote> <p> Simple fluid dynamics on a rotating planet has caught my eye :-) </p> </blockquote> I have collected some material that is needed for this on the page [[Atmospheric and oceanic fluid dynamics]]. However, since I am not very good at programming C++ and also the equations are not that simple, I then decided to try to write a numerical solver for the [[Burgers equation]] in a case where there is also an analytical solution (see that page), one space dimension with periodical boundary conditions. This is far simpler, but is still a nice programming challenge, at least for me. Unfortunately I haven't had any time for this project during the past months, but I hope I can pick it up again during the next months.
  • 55.

    Hi John, Please can you credit Glyn Adgie instead of me for this SR model: he upgraded Allan's javascript.

    Comment Source:Hi John, Please can you credit [[Glyn Adgie]] instead of me for this SR model: he upgraded Allan's javascript.
  • 56.

    For the Code Project, one of the motivations is

    "show how computer code can easily be versioned and published along with any scientific results that are based on the code, by example."

    so would it be safe to say that one of the objectives should be to show why codes are valid/consistent with reality? For example, showing errors of different numerical schemes and computational approaches?

    I ask because I've been blogging about this lately and ended up writing a couple of codes for different schemes to show the possible errors so I was wondering if it could be useful for the Azimuth Code Project as well.

    Comment Source:For the Code Project, one of the motivations is "show how computer code can easily be versioned and published along with any scientific results that are based on the code, by example." so would it be safe to say that one of the objectives should be to show why codes are valid/consistent with reality? For example, showing errors of different numerical schemes and computational approaches? I ask because I've been blogging about this lately and ended up writing a couple of codes for different schemes to show the possible errors so I was wondering if it could be useful for the Azimuth Code Project as well.
  • 57.

    The versioning and data publishing of code aim in practice is satisfied just by using github. IMO assessing how good environmental models are has to be a key Azimuth topic. Numerical approximation has got to be another vital area and I guess programmers here know about this stuff but I'm not sure other people are interested. That's no reason for us not to discuss it :). Can you post a link to your blog?

    Comment Source:The versioning and data publishing of code aim in practice is satisfied just by using github. IMO assessing how good environmental models are has to be a key Azimuth topic. Numerical approximation has got to be another vital area and I guess programmers here know about this stuff but I'm not sure other people are interested. That's no reason for us not to discuss it :). Can you post a link to your blog?
  • 58.
    edited November 2013

    I haven't posted the stability analyses yet because I've only done five less fun schemes right now, but I was thinking of either compiling this into a more cohesive page on the wiki or one nice bit of visual for someone to add to their blog post or something..

    My blog is really more of a way for me to solidify things I'm using, but I hope it will be more useful as I post more!

    Comment Source:I haven't posted the stability analyses yet because I've only done five less fun schemes right now, but I was thinking of either compiling this into a more cohesive page on the wiki or one nice bit of visual for someone to add to their blog post or something.. [My blog](http://plasmanaut.blogspot.com) is really more of a way for me to solidify things I'm using, but I hope it will be more useful as I post more!
  • 59.
    edited April 2014

    I just performed the edit above to Azimuth Code Project, indicated above by Jim a few messages back, to credit Allan and Glyn for the stochastic resonance model.

    Comment Source:I just performed the edit above to [[Azimuth Code Project]], indicated above by Jim a few messages back, to credit Allan and Glyn for the stochastic resonance model.
  • 60.
    edited April 2014

    In this series of messages, I will try to take a fresh look at that Azimuth Code Project page, and review it in light of our current status and objectives.

    1. Statement of Purpose

    From the code project page:

    The Azimuth Code Project is an online open source project that will provide source code for the analysis and simulation software used at the Azimuth project.

    That's good.

    Here is the way that I spontaneously put it in my draft blog on the stochastic resonance model:

    Today we will look at one of the software models from The Azimuth Code Project, which aims to produce educational software that is relevant to the study of climate.

    Very compatible sentences, with the second being more contextual and "green oriented." I like the crispness and generality of the first sentence. Simulations of e.g. evolutionary dynamics also can have an important role within the Azimuth code project, so the climate statement is too restrictive. Maybe these themes can be blended into a new sentence, with the mention of climate in a subordinate clause, such as "...in particular that is relevant to the study of climate."

    I also like the idea of aiming to produce educational software as a proactive goal -- as part of a more general curriculum for science that really matters (nicknamed "Azimuth") -- rather than the code project just being a distribution channel for the source code that is "used at the Azimuth project." The distribution channel statement connotes a higher level of organized activity at Azimuth than actually exists.

    Comment Source:In this series of messages, I will try to take a fresh look at that [[Azimuth Code Project]] page, and review it in light of our current status and objectives. **1. Statement of Purpose** From the code project page: > The Azimuth Code Project is an online open source project that will provide source code for the analysis and simulation software used at the Azimuth project. That's good. Here is the way that I spontaneously put it in my draft blog on the stochastic resonance model: > Today we will look at one of the software models from The Azimuth Code Project, which aims to produce educational software that is relevant to the study of climate. Very compatible sentences, with the second being more contextual and "green oriented." I like the crispness and generality of the first sentence. Simulations of e.g. evolutionary dynamics also can have an important role within the Azimuth code project, so the climate statement is too restrictive. Maybe these themes can be blended into a new sentence, with the mention of climate in a subordinate clause, such as "...in particular that is relevant to the study of climate." I also like the idea of aiming to produce educational software as a proactive goal -- as part of a more general curriculum for science that really matters (nicknamed "Azimuth") -- rather than the code project just being a distribution channel for the source code that is "used at the Azimuth project." The distribution channel statement connotes a higher level of organized activity at Azimuth than actually exists.
  • 61.
    edited April 2014

    2. Idea and Online Models

    Good stuff!

    Comment Source:**2. Idea and Online Models** Good stuff!
  • 62.

    3. Philosophy, Motivation and Objectives

    I’m basically in agreement with the Philosophy section, though I find the the sentence “This needs to change!” to have a more bluntly activist tone than I would prefer on the home page for the Code Project.

    The plot thickens in the section “Motivation and Objectives,” which, though I agree with much of it, reads too much like sermon to me – especially with the rapid succession of boldfaced sentences. Due to this form, the impact of this on a typical programmer may well push them further away from the ideas that are denoted there.

    Let’s face it, we’re a software project that is struggling to find any contributors at all, let alone programmers who subscribe to a given set of ideals. There’s already a Free Software Foundation, and open science initiatives, etc. I just don’t see these goals as part of our true leading edge, which, as I see it, is the relentless pursuit of science that really matters. So open science may be implied by our goals, and we are in fact practicing it, but I don’t see a need to make it part of a group doctrine – some genuine Azimuthers may have different ideas.

    Let me put it another way. On the rest of the Azimuth pages, we gain some credibility by presenting a factual perspective, and avoiding partisanship in our articles. Why should the code project have such a radically different tone?

    Comment Source:**3. Philosophy, Motivation and Objectives** I’m basically in agreement with the Philosophy section, though I find the the sentence “This needs to change!” to have a more bluntly activist tone than I would prefer on the home page for the Code Project. The plot thickens in the section “Motivation and Objectives,” which, though I agree with much of it, reads too much like sermon to me – especially with the rapid succession of boldfaced sentences. Due to this form, the impact of this on a typical programmer may well push them further away from the ideas that are denoted there. Let’s face it, we’re a software project that is struggling to find any contributors at all, let alone programmers who subscribe to a given set of ideals. There’s already a Free Software Foundation, and open science initiatives, etc. I just don’t see these goals as part of our true leading edge, which, as I see it, is the relentless pursuit of science that really matters. So open science may be implied by our goals, and we are in fact practicing it, but I don’t see a need to make it part of a group doctrine – some genuine Azimuthers may have different ideas. Let me put it another way. On the rest of the Azimuth pages, we gain some credibility by presenting a factual perspective, and avoiding partisanship in our articles. Why should the code project have such a radically different tone?
  • 63.
    edited April 2014

    Here are my recommendations for how to handle Philosophy, Motivation and Objectives.

    What I would most prefer is to create a separate page, with a title to the effect of "relationship between Azimuth and the open source and open science movements." (A nicer title, of course!) We can reference this from the Azimuth Code Project page. That will be the place to hash out the "ideological" issues here.

    That aside, I think it's great for us to use the LGPL, and to say that we want to contribute to world's supply of free software, which is a tremendous resource for science in the present era.

    Comment Source:Here are my recommendations for how to handle Philosophy, Motivation and Objectives. What I would most prefer is to create a separate page, with a title to the effect of "relationship between Azimuth and the open source and open science movements." (A nicer title, of course!) We can reference this from the Azimuth Code Project page. That will be the place to hash out the "ideological" issues here. That aside, I think it's great for us to use the LGPL, and to say that we want to contribute to world's supply of free software, which is a tremendous resource for science in the present era.
  • 64.

    4. Technical choices

    This is just a two-sentence stub, which barely says anything.

    Let's nix this section, and recreate it if the need arises later. If anything, our technical choices so far would be concerned with our use of Javascript for interactive climate model programming.

    Comment Source:**4. Technical choices** This is just a two-sentence stub, which barely says anything. Let's nix this section, and recreate it if the need arises later. If anything, our technical choices so far would be concerned with our use of Javascript for interactive climate model programming.
  • 65.

    5. References

    These look a bit sundry: Modelling tool, Visualization tool, and Some physics simulation websites

    The modelling page is just a stub, and is not very cohesive. So we should remove the link from the Code Project page. Separately, let's consider either putting some effort into that page, or deleting it.

    Visualization looks pretty good (from cursory reading).

    Comment Source:**5. References** These look a bit sundry: [[Modelling tool]], [[Visualization tool]], and Some physics simulation websites The modelling page is just a stub, and is not very cohesive. So we should remove the link from the Code Project page. Separately, let's consider either putting some effort into that page, or deleting it. Visualization looks pretty good (from cursory reading).
  • 66.

    6. Coding practice

    The link to the paper by Roedy Green is broken. It is described as "a funny collection of coding practices that you should use, if you don't want anyone else to understand and maintain your code.

    I haven't read the references there, which are not full citations.

    But most of all, do we really need to be giving pointers on coding practices to the programmers who we hope will visit our page? If there are N programmers in the world, then there are already O(N) ideas about coding practices.

    The whole section feels like a tangent from our goals, and I suggest that we nix it.

    Comment Source:**6. Coding practice** The link to the paper by Roedy Green is broken. It is described as "a funny collection of coding practices that you should use, if you don't want anyone else to understand and maintain your code. I haven't read the references there, which are not full citations. But most of all, do we really need to be giving pointers on coding practices to the programmers who we hope will visit our page? If there are N programmers in the world, then there are already O(N) ideas about coding practices. The whole section feels like a tangent from our goals, and I suggest that we nix it.
  • 67.

    I think what might be good would be a "wishlist" of things to program with maybe a couple of more specific search terms/references for anyone to look through.. Of course, that would require a more meta-programming approach, I think. Personally, I would love to write up some more useful interactive codes, but my research projects have been picking up speed lately so I haven't been able to go through enough of the reference modeling book that is cited. It would make it much easier for me (or anyone else looking to contribute) if such a wishlist existed (even in rudimentary form!).

    Comment Source:I think what might be good would be a "wishlist" of things to program with maybe a couple of more specific search terms/references for anyone to look through.. Of course, that would require a more meta-programming approach, I think. Personally, I would love to write up some more useful interactive codes, but my research projects have been picking up speed lately so I haven't been able to go through enough of the reference modeling book that is cited. It would make it much easier for me (or anyone else looking to contribute) if such a wishlist existed (even in rudimentary form!).
  • 68.

    The modelling page is just a stub, and is not very cohesive. So we should remove the link from the Code Project page. Separately, let’s consider either putting some effort into that page, or deleting it.

    I now added some Wikipedia entries to the page. That is not much but at least a start, which I think is eventually important for someone who wants at least to get an overview....including an overview about the ressources of Azimuth.

    Unfortunately the list of modelling tools had been deleted on wikipedia. I couldnt find out why that had been done. Should one eventually think about copying useful Wikipedia entries before they get eventually deleted?

    Comment Source:>The modelling page is just a stub, and is not very cohesive. So we should remove the link from the Code Project page. Separately, let’s consider either putting some effort into that page, or deleting it. I now added some Wikipedia entries to the page. That is not much but at least a start, which I think is eventually important for someone who wants at least to get an overview....including an overview about the ressources of Azimuth. Unfortunately the list of modelling tools had been deleted on wikipedia. I couldnt find out why that had been done. Should one eventually think about copying useful Wikipedia entries before they get eventually deleted?
  • 69.
    edited April 2014

    David wrote:

    I just performed the edit above to Azimuth Code Project, indicated above by Jim a few messages back, to credit Allan and Glyn for the stochastic resonance model.

    Good! Jim should have just edited it himself... and I hope you, too, feel no fear in making edits to improve things. The great thing about a wiki is that we can always look at old versions of the pages. So, if you accidentally go too far and make an edit that someone else dislikes, we can discuss it and fix the problem. This should empower you to be bold.

    I’m basically in agreement with the Philosophy section, though I find the the sentence “This needs to change!” to have a more bluntly activist tone than I would prefer on the home page for the Code Project.

    Change it then! Maybe that statement was self-referential... maybe it needs to change.

    Let me put it another way. On the rest of the Azimuth pages, we gain some credibility by presenting a factual perspective, and avoiding partisanship in our articles. Why should the code project have such a radically different tone?

    If I treat that question as non-rhetorical, the answer is: because Tim van Beek wrote it, and he was feeling upset about the deficiencies in the coding practices of climate scientists. His idea of the Azimuth Coding Project was that it should serve as a role model for how scientific programming should be done. Unfortunately he didn't have the energy to make this dream a reality, even back before he got a management job and disappeared entirely. So: feel free to change the vision!

    In the long run, only those who produce a lot of "content" (blog articles, wiki pages, code, etc.) will get to see their visions realized here.

    What I would most prefer is to create a separate page, with a title to the effect of “relationship between Azimuth and the open source and open science movements.” (A nicer title, of course!) We can reference this from the Azimuth Code Project page. That will be the place to hash out the “ideological” issues here.

    Do it!

    4. Technical choices

    This is just a two-sentence stub, which barely says anything.

    Let’s nix this section, and recreate it if the need arises later.

    Okay, do it!

    The modelling page is just a stub, and is not very cohesive. So we should remove the link from the Code Project page.

    Do it! For small things like this, it may take longer to discuss doing it than to do it.

    I do want to hear what people are doing to the wiki pages. But it's much faster to:

    • do something, then

    • go to the Forum and say "I did X"

    than to

    • go to the Forum, say "I think we should do X", then

    • get a response from me (or not, since I may be distracted), then

    • do it (or not, since you may forget to, especially if I don't respond), and finally

    • go back to the Forum and say "I did X".

    You can probably accomplish at least 2 times as much using the former workflow.

    For big controversial changes of course it's good to discuss things beforehand. But right now you are talking about a page that few people visit &mash; because the Azimuth Code Project is not very active. I hope you energize it to the point where changes in this page become issues of heated discussion! But that's not how it is now.

    Separately, let’s consider either putting some effort into that page, or deleting it.

    Okay, you've finally run into an issue where I want to rein you in. Please don't delete pages except for spammy or terminally stupid pages; it's okay to have stubby pages that lie dormant for years and finally blossom when someone puts some work into them. They don't hurt anyone in the meantime.

    It's also technically rather difficult to make pages truly disappear! Andrew Stacey needs to intervene.

    6. Coding practice

    ... but most of all, do we really need to be giving pointers on coding practices to the programmers who we hope will visit our page? If there are N programmers in the world, then there are already O(N) ideas about coding practices.

    The whole section feels like a tangent from our goals, and I suggest that we nix it.

    Okay, go ahead. This was central to Tim van Beek's vision of Azimuth as a role model for scientific programmers. However, as I mentioned, he never had the energy to make his vision a reality.

    Comment Source:David wrote: > I just performed the edit above to Azimuth Code Project, indicated above by Jim a few messages back, to credit Allan and Glyn for the stochastic resonance model. Good! Jim should have just edited it himself... and I hope you, too, feel no fear in making edits to improve things. The great thing about a wiki is that we can always look at old versions of the pages. So, if you accidentally go too far and make an edit that someone else dislikes, we can discuss it and fix the problem. This should empower you to **be bold**. > I’m basically in agreement with the Philosophy section, though I find the the sentence “This needs to change!” to have a more bluntly activist tone than I would prefer on the home page for the Code Project. Change it then! Maybe that statement was self-referential... maybe it needs to change. <img src = "http://math.ucr.edu/home/baez/emoticons/tongue2.gif" alt = ""/> > Let me put it another way. On the rest of the Azimuth pages, we gain some credibility by presenting a factual perspective, and avoiding partisanship in our articles. Why should the code project have such a radically different tone? If I treat that question as non-rhetorical, the answer is: because Tim van Beek wrote it, and he was feeling upset about the deficiencies in the coding practices of climate scientists. His idea of the Azimuth Coding Project was that it should serve as a role model for how scientific programming should be done. Unfortunately he didn't have the energy to make this dream a reality, even back before he got a management job and disappeared entirely. So: feel free to change the vision! In the long run, only those who produce a lot of "content" (blog articles, wiki pages, code, etc.) will get to see their visions realized here. > What I would most prefer is to create a separate page, with a title to the effect of “relationship between Azimuth and the open source and open science movements.” (A nicer title, of course!) We can reference this from the Azimuth Code Project page. That will be the place to hash out the “ideological” issues here. Do it! > **4. Technical choices** > This is just a two-sentence stub, which barely says anything. > Let’s nix this section, and recreate it if the need arises later. Okay, do it! > The modelling page is just a stub, and is not very cohesive. So we should remove the link from the Code Project page. Do it! For small things like this, it may take longer to discuss doing it than to do it. I do want to hear what people are doing to the wiki pages. But it's much faster to: * do something, then * go to the Forum and say "I did X" than to * go to the Forum, say "I think we should do X", then * get a response from me (or not, since I may be distracted), then * do it (or not, since you may forget to, especially if I don't respond), and finally * go back to the Forum and say "I did X". You can probably accomplish at least 2 times as much using the former workflow. For big controversial changes of course it's good to discuss things beforehand. But right now you are talking about a page that few people visit &mash; because the Azimuth Code Project is not very active. I hope you energize it to the point where changes in this page become issues of heated discussion! But that's not how it is now. > Separately, let’s consider either putting some effort into that page, or deleting it. Okay, you've finally run into an issue where I want to rein you in. Please don't delete pages except for spammy or terminally stupid pages; it's okay to have stubby pages that lie dormant for years and finally blossom when someone puts some work into them. They don't hurt anyone in the meantime. It's also technically rather difficult to make pages truly disappear! Andrew Stacey needs to intervene. > **6. Coding practice** > ... but most of all, do we really need to be giving pointers on coding practices to the programmers who we hope will visit our page? If there are N programmers in the world, then there are already O(N) ideas about coding practices. > The whole section feels like a tangent from our goals, and I suggest that we nix it. Okay, go ahead. This was central to Tim van Beek's vision of Azimuth as a role model for scientific programmers. However, as I mentioned, he never had the energy to make his vision a reality.
  • 70.

    Cool, thanks.

    I started on these changes, deleted coding practice and technical choices.

    Nad, thanks for the new work on the Modelling page. Now it looks like its at the start of a healthy life. So I'll keep the link to it on the code page. Can you take it any further...

    Comment Source:Cool, thanks. I started on these changes, deleted coding practice and technical choices. Nad, thanks for the new work on the Modelling page. Now it looks like its at the start of a healthy life. So I'll keep the link to it on the code page. Can you take it any further...
  • 71.
    edited April 2014

    Calvin wrote:

    I think what might be good would be a “wishlist” of things to program with maybe a couple of more specific search terms/references for anyone to look through

    Calvin, I agree, and have been wishing for a wishlist myself. The fact is, it is not a given right now, and needs to be created and/or re-energized. When I get more up to speed on the code project, I hope to contribute to the wishlist.

    Comment Source:Calvin wrote: > I think what might be good would be a “wishlist” of things to program with maybe a couple of more specific search terms/references for anyone to look through Calvin, I agree, and have been wishing for a wishlist myself. The fact is, it is not a given right now, and needs to be created and/or re-energized. When I get more up to speed on the code project, I hope to contribute to the wishlist.
  • 72.
    edited April 2014

    I joined this project too recently to find out that Tim van Beek had this "role-model" vision - I wondered why there was a redundant sketchy page about standard coding practices.

    I've reposted Tim's point that climate models must have bugs before. This is claimed to be an issue by AGW deniers and needs to be explained. Some of this code is some of the most tested scientific code there is. That doesn't mean that anybody in their right mind would start using Fortran today for pretty much anything.

    I think Azimuth is well placed to carry out interdisciplinary reviews of all sorts of science and software engineering.

    The supposed prototype Azimuth server subscription on Bluehost expires in 2 months. I'm going to ditch it for some better OS, either NixOS or Archlinux in the cloud instead of Centos-6 which has for example a gcc C compiler (gcc-4.4) 3 versions behind the current one (gcc-4.7) and an old glibc which make it unsuitable for development purposes.

    The idea of an Azimuth server, as Stephan Liljegren suggested it would be, was a failure. My guess is that it failed mostly because I didn't produce any working interactive models (although the Snap server worked flawlessly for 2 years). I think that if Azimuth wants to serve up any models a donated cloud service would probably be most appropriate.

    I'm in the middle of a system rebuild thrutch so I might be slow in replying to any comments.

    Great that you restarted this thread Dave :)

    PS John. I think there was also a miscredit on your blog as well which I can't edit.

    Comment Source:I joined this project too recently to find out that [[Tim van Beek]] had this "role-model" vision - I wondered why there was a redundant sketchy page about standard coding practices. I've reposted Tim's point that climate models must have bugs before. This is claimed to be an issue by AGW deniers and needs to be explained. Some of this code is some of the most tested scientific code there is. That doesn't mean that anybody in their right mind would start using Fortran today for pretty much anything. I think Azimuth is well placed to carry out interdisciplinary reviews of all sorts of science and software engineering. The supposed prototype Azimuth server subscription on Bluehost expires in 2 months. I'm going to ditch it for some better OS, either NixOS or Archlinux in the cloud instead of Centos-6 which has for example a gcc C compiler (gcc-4.4) 3 versions behind the current one (gcc-4.7) and an old glibc which make it unsuitable for development purposes. The idea of an Azimuth server, as [[Stephan Liljegren]] suggested it would be, was a failure. My guess is that it failed mostly because I didn't produce any working interactive models (although the Snap server worked flawlessly for 2 years). I think that if Azimuth wants to serve up any models a donated cloud service would probably be most appropriate. I'm in the middle of a system rebuild thrutch so I might be slow in replying to any comments. Great that you restarted this thread Dave :) PS John. I think there was also a miscredit on your blog as well which I can't edit.
  • 73.

    Nad, thanks for the new work on the Modelling page. Now it looks like its at the start of a healthy life. So I’ll keep the link to it on the code page. Can you take it any further…

    I am not sure wether I can do here much - as said I thought it might be

    important for someone who wants at least to get an overview….including an overview about the ressources of Azimuth.

    In particular I haven't done much modelling myself - I was trying blender just a little bit and thus I linked to rather general modelling tools. I think the page might be mostly interesting for aspiring climate modellers if there would be more links to concrete climate modelling tools, but those can probably only be provided by a pro. But then I don't know how these are documented, i.e. wether they are utilizable at all without guidance.

    Comment Source:>Nad, thanks for the new work on the Modelling page. Now it looks like its at the start of a healthy life. So I’ll keep the link to it on the code page. Can you take it any further… I am not sure wether I can do here much - as said I thought it might be > important for someone who wants at least to get an overview….including an overview about the ressources of Azimuth. In particular I haven't done much modelling myself - I was trying blender just a little bit and thus I linked to rather general modelling tools. I think the page might be mostly interesting for aspiring climate modellers if there would be more links to concrete climate modelling tools, but those can probably only be provided by a pro. But then I don't know how these are documented, i.e. wether they are utilizable at all without guidance.
  • 74.

    PS John. I think there was also a miscredit on your blog as well which I can’t edit.

    If you tell me which blog article, I'll be glad to edit it. I have a certain number of blog articles. I can't see your name here:

    Nor here...

    oh, wait. It was here:

    Okay, I've removed it. You are now uncredited.

    Comment Source:> PS John. I think there was also a miscredit on your blog as well which I can’t edit. If you tell me which blog article, I'll be glad to edit it. I have a certain number of blog articles. I can't see your name here: * [Mathematics of the Environment (Part 8)](http://johncarlosbaez.wordpress.com/2012/11/20/mathematics-of-the-environment-part-8/) - describing Michael Knap's version of the program. Nor here... oh, wait. It was here: * [Mathematics of the Environment (Part 8)](http://johncarlosbaez.wordpress.com/2012/11/13/mathematics-of-the-environment-part-7/) - describing a stochastic resonance model. Okay, I've removed it. You are now uncredited.
  • 75.

    Here is one reference that would be needed and appropriate for our current state of affairs: good tutorial and reference material on Javascript. If anyone knows of such, could they reinstate the References section, and add it.

    Comment Source:Here is one reference that would be needed and appropriate for our current state of affairs: good tutorial and reference material on Javascript. If anyone knows of such, could they reinstate the References section, and add it.
  • 76.
    edited April 2014

    nad I edited a couple of pages and pluralised their names to Modelling tools and Visualization tools. I've started 2 surveys of modelling software: one on the Petri net page which John Baez asked me to post and another, way to go, page on Functional reactive programming.

    There is a Javascript implementation of FRP called Flapjax and it would be great if somebody could gave it a try. I noticed on the Haskell cafe mailing list that Allan Erskine is learning Agda-HoTT; Allan, are you about? There's a prototype Agda implementation of FRP obviously called agda-frp. Is there any chance you could look at it?

    I have a draft page on automatic generation of Javascript using the various Haskell Javascript compilers: Fay, Haste and ghcjs which I've been playing with, Ghcjs is now integrated into the Haskell compiler ghc-7.8.1 (bleeding only released this week) and emits reasonable Javascript by passing a flag --ghcjs to the compiler.

    John Baez: Can you decide what to do with the Petri net tools list?. It's still very sketchy but fairly comprehensive.

    Comment Source:[[nad]] I edited a couple of pages and pluralised their names to [[Modelling tools]] and [[Visualization tools]]. I've started 2 surveys of modelling software: one on the [[Petri net]] page which [[John Baez]] asked me to post and another, way to go, page on [[Functional reactive programming]]. There is a Javascript implementation of FRP called Flapjax and it would be great if somebody could gave it a try. I noticed on the Haskell cafe mailing list that [[Allan Erskine]] is learning Agda-HoTT; Allan, are you about? There's a prototype Agda implementation of FRP obviously called agda-frp. Is there any chance you could look at it? I have a draft page on automatic generation of Javascript using the various Haskell Javascript compilers: Fay, Haste and ghcjs which I've been playing with, Ghcjs is now integrated into the Haskell compiler ghc-7.8.1 (bleeding only released this week) and emits reasonable Javascript by passing a flag --ghcjs to the compiler. [[John Baez]]: Can you decide what to do with the Petri net tools list?. It's still very sketchy but fairly comprehensive.
  • 77.
    nad
    edited April 2014

    nad I edited a couple of pages and pluralised their names to Modelling Tools and Visualisation Tools. I’ve started 2 surveys of modelling software: one on the Petri net page which John Baez asked me to post and another, way to go, page on Functional Reactive Programming.

    ????

    what are the Modelling Tools and Visualisation Tools pages? They are empty. Do you want to transfer the content and all the links to the previous ones? The page on Functional Reactive Programming is also empty.

    Comment Source:>nad I edited a couple of pages and pluralised their names to Modelling Tools and Visualisation Tools. I’ve started 2 surveys of modelling software: one on the Petri net page which John Baez asked me to post and another, way to go, page on Functional Reactive Programming. ???? what are the Modelling Tools and Visualisation Tools pages? They are empty. Do you want to transfer the content and all the links to the previous ones? The page on Functional Reactive Programming is also empty.
  • 78.
    edited April 2014

    Sorry nad, forgot to lowercase the the second word of the names. Fixed. I think you were spot on with your selections of InsightMaker and NetLogo. I can't find many other packages which can be used so easily without any programming. I might have wasted time listing several dozen modelling packages on the Petri net and FRP pages but I don't want to repeat existing stuff. Now I need to code up some demos as evidence for tools we might use to get a lot of interactive animated PDE etc. solutions online.

    Comment Source:Sorry nad, forgot to lowercase the the second word of the names. Fixed. I think you were spot on with your selections of InsightMaker and NetLogo. I can't find many other packages which can be used so easily without any programming. I might have wasted time listing several dozen modelling packages on the Petri net and FRP pages but I don't want to repeat existing stuff. Now I need to code up some demos as evidence for tools we might use to get a lot of interactive animated PDE etc. solutions online.
  • 79.
    nad
    edited April 2014

    Sorry nad, forgot to lowercase the the second word of the names. Fixed.

    sorry, still can't find the visualization tools.

    I had a quick glance at the beginning of your FRP page. You mention some modelling tools, which could also be mentioned on the modelling tools page.

    Didn't understand that:

    A non-causal model is a system of differential algebraic equations (DAEs).

    that is DAEs seem to have still a distinct arrow of time. ????

    Also didn't quite get why one should introduce the name "behaviour" to time-dependend continous(?) function, similarily for event.

    Comment Source:>Sorry nad, forgot to lowercase the the second word of the names. Fixed. sorry, still can't find the visualization tools. I had a quick glance at the beginning of your FRP page. You mention some modelling tools, which could also be mentioned on the modelling tools page. Didn't understand that: >A non-causal model is a system of differential algebraic equations (DAEs). that is <a href="http://en.wikipedia.org/wiki/Differential_algebraic_equation">DAEs</a> seem to have still a distinct arrow of time. ???? Also didn't quite get why one should introduce the name "behaviour" to time-dependend continous(?) function, similarily for event.
  • 80.
    edited April 2014

    Corrected Visualization tools. A non-causal model doesn't specify direction of causality but still happens over time. Behaviours and events just correspond to continuous and discrete.

    Comment Source:Corrected [[Visualization tools]]. A non-causal model doesn't specify direction of causality but still happens over time. Behaviours and events just correspond to continuous and discrete.
  • 81.

    Jim wrote

    John Baez: Can you decide what to do with the Petri net tools list? It’s still very sketchy but fairly comprehensive.

    Thanks a lot for doing this!

    When I get back to work on the book I'll write a chapter or appendix on this subject. I will probably have questions then... but let me take a look now and see what I think.

    Comment Source:Jim wrote > John Baez: Can you decide what to do with the Petri net tools list? It’s still very sketchy but fairly comprehensive. Thanks a lot for doing this! <img src = "http://math.ucr.edu/home/baez/emoticons/thumbsup.gif" alt = ""/> When I get back to work on the book I'll write a chapter or appendix on this subject. I will probably have questions then... but let me take a look now and see what I think.
  • 82.
    nad
    edited April 2014

    Jim wrote:

    They’re called behaviours and events because that’s what systems do and that’s what we want to model :).

    you can of course name things as ever you want. There is a story by Peter Bichsel which thematizises renaming. Its in german but with google translate its probably still readable.

    ->ein tisch ist ein tisch by Peter Bichsel (in german)

    Comment Source:Jim wrote: >They’re called behaviours and events because that’s what systems do and that’s what we want to model :). you can of course name things as ever you want. There is a story by Peter Bichsel which thematizises renaming. Its in german but with google translate its probably still readable. -><a href="http://www.mondeberbere.com/taddangiwin/200411_ein_tisch.htm">ein tisch ist ein tisch by Peter Bichsel (in german)</a>
  • 83.

    Okay, I took a look! It looks good! I'm planning to write something that's mainly links to software that runs Petri nets, with short sketchy comments, mainly just enough to give people an idea of the differences between the various products. So, we don't need a lot more... though again, I'll probably have some questions when I get around to actually writing about this.

    I'll show you a draft of what I write, and you may have suggestions then.

    Comment Source:Okay, I took a look! It looks good! I'm planning to write something that's mainly links to software that runs Petri nets, with short sketchy comments, mainly just enough to give people an idea of the differences between the various products. So, we don't need a lot more... though again, I'll probably have some questions when I get around to actually writing about this. I'll show you a draft of what I write, and you may have suggestions then.
  • 84.

    Nad and Dave,

    As you're organising the software pages of the wiki I'll leave it up to you to link to, move or delete whatever's on the Petri net and Functional reactive programming pages. Thanks for taking this on.

    There are a number of modelling in the Cloud services where people don't have to do all that horrible downloading and installing rigmarole. Would "Modelling in the Cloud" or some such also be a useful page. Also I don't quite know where reviews on numerical method packages should go.

    Cheers

    PS. John, I've just found another miscredit for Allan, Glyn and Michael's code. It's item '54 above on this page. I hate to waste your time being picky but being unjustifiably credited with other people's work is embarrasing. Cheers.

    Comment Source:Nad and Dave, As you're organising the software pages of the wiki I'll leave it up to you to link to, move or delete whatever's on the [[Petri net]] and [[Functional reactive programming]] pages. Thanks for taking this on. There are a number of modelling in the Cloud services where people don't have to do all that horrible downloading and installing rigmarole. Would "Modelling in the Cloud" or some such also be a useful page. Also I don't quite know where reviews on numerical method packages should go. Cheers PS. John, I've just found another miscredit for Allan, Glyn and Michael's code. It's item '54 above on this page. I hate to waste your time being picky but being unjustifiably credited with other people's work is embarrasing. Cheers.
  • 85.

    As you’re organising the software pages of the wiki I’ll leave it up to you to link to, move or delete whatever’s on the Petri net and Functional reactive programming pages.

    I am not organizing the software pages of the wiki. I helped Tim edit the visualization page, gave my opinion on the existence of modelling tools page and made some remarks on the FRP page.

    Comment Source:>As you’re organising the software pages of the wiki I’ll leave it up to you to link to, move or delete whatever’s on the Petri net and Functional reactive programming pages. I am not organizing the software pages of the wiki. I helped Tim edit the visualization page, gave my opinion on the existence of modelling tools page and made some remarks on the FRP page.
  • 86.

    Nad, I was rushed. I've re-edited item '81 above, it's just discrete + continuous = hybrid, as described in George Geordidge's Hydra thesis.

    John. Thanks for the appreciative smileys :). I hope I'll have some better informed opinions whenever you get to that part of the book

    Comment Source:Nad, I was rushed. I've re-edited item '81 above, it's just discrete + continuous = hybrid, as described in George Geordidge's Hydra thesis. John. Thanks for the appreciative smileys :). I hope I'll have some better informed opinions whenever you get to that part of the book
  • 87.
    edited May 2014

    In preparation for any new visitors who might stop by after my new blogs on the Azimuth Code Project, I did some organizing of the page.

    • New introduction, which is a fusion of what was there before, and parts of my introduction to Blog - The stochastic resonance program (part 1)

    • Moved the section called "Motivation and Objectives" to a new page, Azimuth Code Project charter (draft statement), and added a link to it the Philosophy section.

    • Collected various scattered references into the one References section at the end.

    • Removed link to software engineering in climate science, which was a stub. That page did contain a couple of good references (to Steve Easterbrook's blog, and to a book on the computing infrastructure for climate science), which added to the References section of the code project page.

    • Removed link to semantic web applications and environment, which was a stub that pointed to a draft of Nad's blog article. In its place, I put a link to the final, arXiv version of this article, in the References section.

    • Created a section called Source Code Repositories:

    We currently have two repositories:

    Azimuth Project Homepage (SVN), Google project hosting.

    Azimuth Project 2 Git, Google project hosting.

    Caveat: the repositories are not populated with much of our present code, which is spread across the web. We need to take an inventory and get our stuff organized. Perhaps we should migrate everything over to the GIT repository and delete the SVN repository.

    Jim, have I made any misstatement here? The GIT repository looks essentially empty (except for some C++ package classes that Tim checked in), and I didn't see the javascript models in the SVN repository -- maybe I didn't look in the right place. I'd like people to know that we have these repositories available, but to keep their expectations low about what contents might be in there at present -- so they don't get disappointed by an empty GIT repository. Also, I proposed that we move everything over to GIT, because having two repositories sounds like a formula for trouble and confusion.

    Comment Source:In preparation for any new visitors who might stop by after my new blogs on the Azimuth Code Project, I did some organizing of the page. * New introduction, which is a fusion of what was there before, and parts of my introduction to [[Blog - The stochastic resonance program (part 1)]] * Moved the section called "Motivation and Objectives" to a new page, [[Azimuth Code Project charter (draft statement)]], and added a link to it the Philosophy section. * Collected various scattered references into the one References section at the end. * Removed link to [[software engineering in climate science]], which was a stub. That page did contain a couple of good references (to Steve Easterbrook's blog, and to a book on the computing infrastructure for climate science), which added to the References section of the code project page. * Removed link to [[semantic web applications and environment]], which was a stub that pointed to a draft of Nad's blog article. In its place, I put a link to the final, arXiv version of this article, in the References section. * Created a section called Source Code Repositories: > We currently have two repositories: > &bull; [Azimuth Project Homepage (SVN)](http://code.google.com/p/azimuthproject/), Google project hosting. > &bull; [Azimuth Project 2 Git](http://code.google.com/p/azimuthproject-2/), Google project hosting. > Caveat: the repositories are not populated with much of our present code, which is spread across the web. We need to take an inventory and get our stuff organized. Perhaps we should migrate everything over to the GIT repository and delete the SVN repository. Jim, have I made any misstatement here? The GIT repository looks essentially empty (except for some C++ package classes that Tim checked in), and I didn't see the javascript models in the SVN repository -- maybe I didn't look in the right place. I'd like people to know that we have these repositories available, but to keep their expectations low about what contents might be in there at present -- so they don't get disappointed by an empty GIT repository. Also, I proposed that we move everything over to GIT, because having two repositories sounds like a formula for trouble and confusion.
  • 88.
    edited May 2014

    Hi Dave,

    Tim van Beek set up an initial svn repo with all the code for different projects in a single download. After some discussion Allan Erskine started a new azimuth-project repo on github. I couldn't find the site via the github search box for some reason.

    If you've got a github account and you give me your user name I think I can add you to the list of committers.

    I definitely vote for moving all existing Azimuth code examples to there. I think it might help avoid confusion if Tim was prepared to destroy the googlecode repos but I suppose we could email him to ask when we've ripped the existing code.

    Comment Source:Hi Dave, [[Tim van Beek]] set up an initial svn repo with all the code for different projects in a single download. After some discussion [[Allan Erskine]] started a new [azimuth-project repo]([https://github.com/azimuth-project) on github. I couldn't find the site via the github search box for some reason. If you've got a github account and you give me your user name I think I can add you to the list of committers. I definitely vote for moving all existing Azimuth code examples to there. I think it might help avoid confusion if Tim was prepared to destroy the googlecode repos but I suppose we could email him to ask when we've ripped the existing code.
  • 89.

    Hi Jim, I just created a github account, davidtanzer.

    This is the repo that I got to from the Azimuth Code Project page: azimuthproject-2. That would be great if you could add me to the committer list.

    Thanks

    Comment Source:Hi Jim, I just created a github account, davidtanzer. This is the repo that I got to from the Azimuth Code Project page: [azimuthproject-2](http://code.google.com/p/azimuthproject-2/). That would be great if you could add me to the committer list. Thanks
  • 90.

    Willco! Glyn, our mate Daniel Everett (C++ edu whiteboard programer), and I modified [[David Tweed]'s C++ SDE code from the googlecode svn repo to build with cmake. I uploaded it some where currently unkown but I guess I could find it. I also managed to build (good old reliable jvm) Tim's bistable maven java code. hth.

    Comment Source:Willco! Glyn, our mate Daniel Everett (C++ edu whiteboard programer), and I modified [[David Tweed]'s C++ SDE code from the googlecode svn repo to build with cmake. I uploaded it some where currently unkown but I guess I could find it. I also managed to build (good old reliable jvm) Tim's bistable maven java code. hth.
  • 91.

    Removed link to semantic web applications and environment, which was a stub that pointed to a draft of Nad’s blog article. In its place, I put a link to the final, arXiv version of this article, in the References section.

    David thanks for cleaning up, but I think this link should be kept. That is it is to be expected that there are and will be a couple of big data applications linked to environmental issues and so semantic applications could be regarded as a subject in its own, just like visualization and modelling. As I said for the modelling - I think it is useful to display that one is aware of a subject and it is useful to point to it even if untended, visitors of the Azimuth code page would probably want to see what kind of software Azimutheres think of/ is in use here and where there is a lot to do. So I think it is useful to keep "construction sites" . I supplemented the text with respect to this. Furthermore I reinstalled the link, wrote that it contains sofar only a link to one article and made a direct link to the archiv article, which can be removed if there is more in that section.

    Comment Source:>Removed link to semantic web applications and environment, which was a stub that pointed to a draft of Nad’s blog article. In its place, I put a link to the final, arXiv version of this article, in the References section. David thanks for cleaning up, but I think this link should be kept. That is it is to be expected that there are and will be a couple of big data applications linked to environmental issues and so semantic applications could be regarded as a subject in its own, just like visualization and modelling. As I said for the modelling - I think it is useful to display that one is aware of a subject and it is useful to point to it even if untended, visitors of the Azimuth code page would probably want to see what kind of software Azimutheres think of/ is in use here and where there is a lot to do. So I think it is useful to keep "construction sites" . I supplemented the text with respect to this. Furthermore I reinstalled the link, wrote that it contains sofar only a link to one article and made a direct link to the archiv article, which can be removed if there is more in that section.
  • 92.

    Have added davidtanzer to azimuth-project owners on github https://github.com/azimuth-project. Logging on to github.com and typing azimuth-project into the search all repos box doesn't produce any result. This will make it hard for others to find it unless they use the full uri. Will somebody please check if they have the same problem; perhaps there are some public visibility settings.

    Comment Source:Have added davidtanzer to azimuth-project owners on github https://github.com/azimuth-project. Logging on to github.com and typing azimuth-project into the search all repos box doesn't produce any result. This will make it hard for others to find it unless they use the full uri. Will somebody please check if they have the same problem; perhaps there are some public visibility settings.
  • 93.
    edited May 2014

    Dave, I've just pushed a first set of Javascript FRP demos (generated by the Elm and Fay Haskell Javascript compilers) to my github repo. May be some of them should be on the azimuth-project repo? I'm hoping somebody will have a go with flapjax. I think nad might have looked at it? Cheers

    Comment Source:Dave, I've just pushed a first set of Javascript FRP demos (generated by the Elm and Fay Haskell Javascript compilers) to [my github repo](http://github.com/jimstutt/). May be some of them should be on the azimuth-project repo? I'm hoping somebody will have a go with [flapjax](http://www.flapjax-lang.org/). I think [[nad]] might have looked at it? Cheers
  • 94.

    typing azimuth-project into the search all repos box doesn’t produce any result.

    Same here. On the other hand, searching on coalbedo worked.

    Comment Source:> typing azimuth-project into the search all repos box doesn’t produce any result. Same here. On the other hand, searching on coalbedo worked.
  • 95.

    Perhaps its something to do with azimuth-project being an "organisation account" whatever difference that makes. Perhaps it isn't a repo?

    Comment Source:Perhaps its something to do with azimuth-project being an "organisation account" whatever difference that makes. Perhaps it isn't a repo?
  • 96.
    edited May 2014

    Jim thanks for adding me to the list.

    Let's take a bit of time to plan the structure, before checking anything in.

    We now have:

    • google code repository, svn

    • google code repository, git

    • github account, containing two repositories: coalbedo-example, deploy-web

    And the Azimuth Code Project page links to the two google code repositories, but not to github.

    Cleanup time.

    Here are my ideas about how to proceed.

    So we'll go with github. Let's copy anything meaningful from the other repos there, and then if we can get consensus from all interested parties, delete the google code repos. If we can't get that consensus, we can at least post a notice saying that they are obsolete, with a reference to the github repo. And of course change the Azimuth Code Project page to point just to github.

    I think that we should have one main repository on github, for source code, under the azimuth account, with a folder structure that represents our concerns and interests. Since it will contain hand-crafted programs, and README files, etc., it shouldn't get that big, and so downloading all will be no issue.

    If we need this: any large datasets should go into their own repositories.

    If we need this: we can have another repository for larger packages that we might use, written by others. Compilers, class libraries, or what have you. Small "packages," say consisting of a few source files, can be included in our main source repository, in a way that the folder structure will show that it is separate from the Azimuth project.

    The names of these repos could be:

    azimuth-project-source

    azimuth-project-lib

    azimuth-project-dataset1

    azimuth-project-dataset2

    ...

    In a followup message I'll post my ideas for how to organize the folder structure.

    Comment Source:Jim thanks for adding me to the list. Let's take a bit of time to plan the structure, before checking anything in. We now have: * google code repository, svn * google code repository, git * github account, containing two repositories: coalbedo-example, deploy-web And the Azimuth Code Project page links to the two google code repositories, but not to github. Cleanup time. Here are my ideas about how to proceed. So we'll go with github. Let's copy anything meaningful from the other repos there, and then if we can get consensus from all interested parties, delete the google code repos. If we can't get that consensus, we can at least post a notice saying that they are obsolete, with a reference to the github repo. And of course change the Azimuth Code Project page to point just to github. I think that we should have one main repository on github, for _source code_, under the azimuth account, with a folder structure that represents our concerns and interests. Since it will contain hand-crafted programs, and README files, etc., it shouldn't get _that_ big, and so downloading all will be no issue. If we need this: any large datasets should go into their own repositories. If we need this: we can have another repository for larger packages that we might use, written by others. Compilers, class libraries, or what have you. Small "packages," say consisting of a few source files, can be included in our main source repository, in a way that the folder structure will show that it is separate from the Azimuth project. The names of these repos could be: azimuth-project-source azimuth-project-lib azimuth-project-dataset1 azimuth-project-dataset2 ... In a followup message I'll post my ideas for how to organize the folder structure.
  • 97.
    edited May 2014

    Here are my ideas for the folder structure for azimuth-project-source:


    online-models stochastic resonance example ... stochastic energy balance model (Michael Knapp) stochastic energy balance model (Taylor Baldwin) models-in-progress TODO list delayed oscillator ... petri net simulators (if we get to the point of writing our own) ... platforms Javascript code samples, tips JSXGraph functional reactive programming (elm) lib the C++ classes Tim checked into the svn repo ... utilities deploy-web ...

    The structure should "keep our eye on the prize," which is to distribute the source code for the models that we have published, and to encourage and facilitate the creation of new models (taken in a broad sense, to include e.g. Petri net simulators).

    lib is a kitchen sink container. Logically speaking it belongs in the repo azimuth-project-lib, because it is unbounded in size. But judiciously chosen, small elements can be brought into the source repo, if they really stand a chance of helping us to create new models.

    So that's my idea for how to approach it. What do you guys think?

    Comment Source:Here are my ideas for the folder structure for azimuth-project-source: ~~~ online-models stochastic resonance example ... stochastic energy balance model (Michael Knapp) stochastic energy balance model (Taylor Baldwin) models-in-progress TODO list delayed oscillator ... petri net simulators (if we get to the point of writing our own) ... platforms Javascript code samples, tips JSXGraph functional reactive programming (elm) lib the C++ classes Tim checked into the svn repo ... utilities deploy-web ... ~~~ The structure should "keep our eye on the prize," which is to distribute the source code for the models that we have published, and to encourage and facilitate the creation of new models (taken in a broad sense, to include e.g. Petri net simulators). lib is a kitchen sink container. Logically speaking it belongs in the repo azimuth-project-lib, because it is unbounded in size. But judiciously chosen, small elements can be brought into the source repo, if they really stand a chance of helping us to create new models. So that's my idea for how to approach it. What do you guys think?
  • 98.

    By the way that FRP stuff looks really cool, thanks Jim for identifying it.

    Comment Source:By the way that FRP stuff looks really cool, thanks Jim for identifying it.
  • 99.

    Tim classified code by language. In an industrial setting (and something I regret that Haskell's hackage repo hasn't done) it's standard to have dev, test and prod or stable repos. Your schema fits with this dev and stable distinction.

    Allan Erskine proposed using a continuous build system. He suggested Jenkins - I'd go for Travis although I've never used either. I did do round tripping using OCL in Rational Rose (Java) a long time ago and that is the nearest I've seen to diagram-driven programming and imo a very good thing. It's just the next logical step up from VCS.

    I take it as standard that a repo is created for each application. As we have so few the simplest structure is just a flat list of repos so people can clone just the model they want. I think their names should be self-explanatory and no classifying structure is needed.

    I've extracted the R, CPlusPlus, Java, Scala etc. applications from the google code original (svn) repo and chopped off the unneeded parent directories.

    I have them ready to push as individual model repos eg. DiscreteEvenSimulation (CPP), Bistable.. and the others you mention.

    Comment Source:Tim classified code by language. In an industrial setting (and something I regret that Haskell's hackage repo hasn't done) it's standard to have dev, test and prod or stable repos. Your schema fits with this dev and stable distinction. [[Allan Erskine]] proposed using a continuous build system. He suggested Jenkins - I'd go for Travis although I've never used either. I did do round tripping using OCL in Rational Rose (Java) a long time ago and that is the nearest I've seen to diagram-driven programming and imo a very good thing. It's just the next logical step up from VCS. I take it as standard that a repo is created for each application. As we have so few the simplest structure is just a flat list of repos so people can clone just the model they want. I think their names should be self-explanatory and no classifying structure is needed. I've extracted the R, CPlusPlus, Java, Scala etc. applications from the google code original (svn) repo and chopped off the unneeded parent directories. I have them ready to push as individual model repos eg. DiscreteEvenSimulation (CPP), Bistable.. and the others you mention.
  • 100.

    A flat list of repositories, one per application, seems best to me, at least for starters.

    My guess is that most code will belong to two major categories: educational, and experimental. The former will be about making well-known science accessible, the latter whatever happens to interest the programmer. But that is just a guess, and I wouldn't want to impose anything until there is more code.

    Comment Source:A flat list of repositories, one per application, seems best to me, at least for starters. My guess is that most code will belong to two major categories: educational, and experimental. The former will be about making well-known science accessible, the latter whatever happens to interest the programmer. But that is just a guess, and I wouldn't want to impose anything until there is more code.
Sign In or Register to comment.