#### Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Options

# Week308

It would be great if you folks could look at week308 before I officially announce its existence tomorrow, and tell me - here - if it needs to be improved.

Tim van Beek did a lot of the work on this one, so I'm especially interested in his opinion. But others of you may have ideas — especially since it describes the Azimuth Code Project, which is aimed at people who program. Maybe I could do a better job of selling it than I did. I don't want to oversell it, since it's just getting started, but still...

I feel sort of bad that I did not include Graham Jones' work on predator-prey models in my discussion of the Hopf bifurcation. However, I feel I often talk about too many things in a single issue of This Week's Finds, straining the patience of most readers. So, I think it will be better if I do a separate issue on quantitative ecology. This would make a nice example of how the same patterns - limit cycles, etc. - show up in different systems.

Graham: would it be okay if I use your pictures in a future issue of This Week's Finds? I sort of hint that I may do this, near the end of week308.

The day after tomorrow (Christmas) I will go to Vietnam, and I'll probably be out of contact until I get back on January 5th.

• Options
1.

I cannot comment on the math but I think its a good idea to stick to one idea/thread/innovation at the time. because people have different backgrounds. What triggers me with Hopf is that looks simple and accurate (like any of the simple attractors) and then there is many ways of detecting them in time series (doyne farmer formerly Santa Fe Institute) . They write about it in Lui Lam (ed) intro to non-linear physics - ch 6-7. and maybe also later prediction , reduced noise and what not. Or am I missing something?

Comment Source:I cannot comment on the math but I think its a good idea to stick to one idea/thread/innovation at the time. because people have different backgrounds. What triggers me with Hopf is that looks simple and accurate (like any of the simple attractors) and then there is many ways of detecting them in time series (doyne farmer formerly Santa Fe Institute) . They write about it in Lui Lam (ed) intro to non-linear physics - ch 6-7. and maybe also later prediction , reduced noise and what not. Or am I missing something?
• Options
2.
edited December 2010

Thanks, Stafffan. I don't know Lui Lam's book, but I could try to peek at it on Google, or something.

I was hoping you'd get involved in the Azimuth Code Project, but so far you seem more interested in writing pages for the Azimuth Project. That's great too!

Comment Source:Thanks, Stafffan. I don't know Lui Lam's book, but I could try to peek at it on Google, or something. I was hoping you'd get involved in the Azimuth Code Project, but so far you seem more interested in writing pages for the Azimuth Project. That's great too!
• Options
3.

Week 308 looks fine to me, and feel free to use my pictures as and when it suits you!

Comment Source:Week 308 looks fine to me, and feel free to use my pictures as and when it suits you!
• Options
4.
edited December 2010

Hi John,

I think it looks nice (like all the weeks). I would like to make some very small comments — since you were asking for them:

to have an online free access repository and make all software open source

shouldn't there be a link to a page with the software that was used for week 308 too? Or is it already accessible via Azimuth code project? I didn't find it there.

Of course it doesn't prove that our simple model is "right"

doesn't it sound a little strong? You haven't discussed yet what "$r$" is supposed to be according to your model if "$x$" is the SOI. (ok, maybe I should just dig into the references to see what it's supposed to be)

Merry christmas!

Comment Source:Hi John, I think it looks nice (like all the weeks). I would like to make some very small comments &mdash; since you were asking for them: > to have an online free access repository and make all software open source shouldn't there be a link to a page with the software that was used for week 308 too? Or is it already accessible via [[Azimuth code project]]? I didn't find it there. > Of course it doesn't prove that our simple model is "right" doesn't it sound a little strong? You haven't discussed yet what "$r$" is supposed to be according to your model if "$x$" is the SOI. (ok, maybe I should just dig into the references to see what it's supposed to be) Merry christmas!
• Options
5.

that was my plan but you lured me into SEWTHA :-) but seriously i will focus on code at least with one hand the ISBN for the book is 038740614X

I ll add 308 tools now

Comment Source:that was my plan but you lured me into [[SEWTHA]] :-) but seriously i will focus on code at least with one hand the ISBN for the book is 038740614X I ll add 308 tools now
• Options
6.

http://www.azimuthproject.org/azimuth/show/Visualization+tool lists gnuplot already, but its in category methodology. ill put it there i think. i can change category to software. But first I need to have someone refresh my memory on category policy: can one page be in several categories?

Comment Source:[http://www.azimuthproject.org/azimuth/show/Visualization+tool](http://www.azimuthproject.org/azimuth/show/Visualization+tool) lists gnuplot already, but its in category methodology. ill put it there i think. i can change category to software. But first I need to have someone refresh my memory on category policy: can one page be in several categories?
• Options
7.
edited December 2010

Yes you have that on some pages so I'll do that for now. I love having monologues with myself, then i get right answers!

Comment Source:Yes you have that on some pages so I'll do that for now. I love having monologues with myself, then i get right answers!
• Options
8.
edited December 2010

I couldn't wait for Tim anymore, since I'm leaving tomorrow and I want to be around for at least some of the feedback: I put "week308" on the blog.

Thanks for the suggestions, Frederik! I implemented some of them. I still need to add info on what the two variables actually might be in the ENSO cycle.

Comment Source:I couldn't wait for Tim anymore, since I'm leaving tomorrow and I want to be around for at least some of the feedback: I put "week308" [on the blog](http://johncarlosbaez.wordpress.com/2010/12/24/this-weeks-finds-week-308/). Thanks for the suggestions, Frederik! I implemented some of them. I still need to add info on what the two variables actually _might be_ in the ENSO cycle.
• Options
9.

Whoops, sorry, I read week 308 during the draft stage and did not have any substantial criticism - and also no time to check the forum in the days before Christmas...

Comment Source:Whoops, sorry, I read week 308 during the draft stage and did not have any substantial criticism - and also no time to check the forum in the days before Christmas...
• Options
10.

Frederik wrote:

...shouldn't there be a link to a page with the software that was used for week 308 too? Or is it already accessible via Azimuth code project? I didn't find it there.

the code is there :-) but you're right, I need to set up a packaging and distribution mechanism next. I'm taken by surprise by John's interest and promotion of the project, I expected to use it just as a SVN repository for myself for the next few months...

Comment Source:Frederik wrote: <blockquote> <p> ...shouldn't there be a link to a page with the software that was used for week 308 too? Or is it already accessible via Azimuth code project? I didn't find it there. </p> </blockquote> the code is there :-) but you're right, I need to set up a packaging and distribution mechanism next. I'm taken by surprise by John's interest and promotion of the project, I expected to use it just as a SVN repository for myself for the next few months...
• Options
11.

Tim wrote:

I'm taken by surprise by John's interest and promotion of the project...

I did warn you.

I think it's a great idea, and I hope to lure more people into contributing.

When it comes to things like the Azimuth Project wiki, the Azimuth code project, etc., I can never decide when is the best time to promote them. Should I wait until an impressive amount has already been done - slowly, by a small team of people? Or should I announce them as soon as possible? Or should I announce them and keep announcing them? What is the best way to get lots of people involved in these projects? There are probably books on this topic, but I haven't read them.

Comment Source:Tim wrote: > I'm taken by surprise by John's interest and promotion of the project... I did warn you. <img src = "http://math.ucr.edu/home/baez/emoticons/tongue2.gif" alt = ""/> I think it's a great idea, and I hope to lure more people into contributing. When it comes to things like the Azimuth Project wiki, the Azimuth code project, etc., I can never decide when is the best time to promote them. Should I wait until an impressive amount has already been done - slowly, by a small team of people? Or should I announce them as soon as possible? Or should I announce them and keep announcing them? What is the best way to get lots of people involved in these projects? There are probably books on this topic, but I haven't read them.
• Options
12.
edited January 2011

I think the one key point that open source developers who write about things seem to stress is that one key point on a project is the point at which you've "done enough work" to make it concrete what your goals and approach will be. If you attract interest after that, when someone starts arguing for some slightly different goal you can say "that might be a good idea, but it's not what we're doing here". In contrast, projects that get strong interest at the very early conceptual stage often end up with endless discussions about what's going to be done and never get beyond that.

(This is not particularly relevant for me since I'm very unlikely to do any programming on models for climate change; if I do any programming on Azimuth it's likely to be in modelling mitigation/adaptation type stuff, eg., smart-grid, etc.)

Comment Source:I think the one key point that open source developers who write about things seem to stress is that one key point on a project is the point at which you've "done enough work" to make it concrete what your goals and approach will be. If you attract interest after that, when someone starts arguing for some slightly different goal you can say "that might be a good idea, but it's not what we're doing **here**". In contrast, projects that get strong interest at the very early conceptual stage often end up with endless discussions about what's going to be done and never get beyond that. (This is not particularly relevant for me since I'm very unlikely to do any programming on models for climate change; if I do any programming on Azimuth it's likely to be in modelling mitigation/adaptation type stuff, eg., smart-grid, etc.)
• Options
13.

There are books about open source projects by people who have run very successful ones, like SVN itself. There is no "correct way" to promote them, they are usually started by a developer who has a need and does not find a solution, so he creates one himself. Then others find out, mostly by chance, that this solution is of interest to them, too. If they get interested enough in using it, but find it lacking, some of them will engage in the project to add the features they need, or to fix the bugs that they don't like. In this stage, it is important to think about users who are not developers working on the project itself. There is a lot of infrastructure that needs to be set up, documentation, readmes, faq, downloadable and installable packages, all the things that every developer already can do by himself, and therefore has no interest in it, but no external user can. So, that's basically what I have to do next: Think up a deploying strategie and write some documentation :-)

Comment Source:There are books about open source projects by people who have run very successful ones, like SVN itself. There is no "correct way" to promote them, they are usually started by a developer who has a need and does not find a solution, so he creates one himself. Then others find out, mostly by chance, that this solution is of interest to them, too. If they get interested enough in using it, but find it lacking, some of them will engage in the project to add the features they need, or to fix the bugs that they don't like. In this stage, it is important to think about users who are not developers working on the project itself. There is a lot of infrastructure that needs to be set up, documentation, readmes, faq, downloadable and installable packages, all the things that every developer already can do by himself, and therefore has no interest in it, but no external user can. So, that's basically what I have to do next: Think up a deploying strategie and write some documentation :-)