Help make Minsky easier to use

• Tweet

• Tweet

I’ve just put A$10,000 of my own mon­ey towards improv­ing Min­sky, the Open Source pro­gram I have designed to enable econ­o­mists to cre­ate dynam­ic and mon­e­tary mod­els of the econ­o­my. If you sup­port the work I’m doing to help eco­nom­ics escape its 19th cen­tu­ry equi­lib­ri­um fetish, please con­sid­er also mak­ing a dona­tion to Min­sky’s devel­op­ment via Dr Rus­sell Stan­dish’s Pay­Pal account (or via direct deb­it, using the account details you’ll find below):

Min­sky has been pro­grammed almost exclu­sive­ly by Dr Rus­sell Stan­dish, and $10,000 will buy 100 hours of Rus­sel­l’s pro­gram­ming time. About A$230,000 has been spent on it so far–with US$128,000 com­ing from an ini­tial INET grant (when the US$ was worth less than the A$), US$78,000 from a Kick­starter cam­paign, and sundry oth­er amounts from sup­port­ers like Bruce Ram­say, who runs the End­ing Over­lend­ing page that is linked to from this blog. This fund­ing enabled Rus­sell to build the basic func­tion­al­i­ty Min­sky need­ed, along with a lot of inno­v­a­tive “smarts” that set it apart from its much more estab­lished rivals in sys­tem dynamics–programs like Mat­lab’s Simulink, Ven­sim, Stel­la and Vis­sim that cost thou­sands of dol­lars a copy and have been around for decades.

For exam­ple, Min­sky is the only sys­tem dynam­ics pro­gram that lets you use Greek char­ac­ters and sym­bols, super­scripts and sub­scripts; it runs plots dynam­i­cal­ly while a sim­u­la­tion is run­ning (which only Vis­sim also does in the sys­tem dynam­ics prod­uct space), it’s the only pro­gram that lets you insert vari­ables and oper­a­tors by typ­ing direct­ly onto the can­vas rather than hav­ing to use the mouse and tool­box palettes; and of course it’s the only pro­gram that sup­ports dou­ble-entry book­keep­ing to allow com­plex inter-relat­ed finan­cial accounts to be sim­u­lat­ed dynam­i­cal­ly.

But the pro­gram is still incom­plete. Some basic things like an IF/THEN/ELSE block are miss­ing; some aspects of group­ing don’t work prop­er­ly yet, you can’t save part of a Min­sky file as a toolk­it, and so on. I’m putting $10,000 of my own cash in to get these things done now–and there are many oth­er fea­tures that should be added. These range from sim­ple things like adding short­cut keys for “Save As” to the final ambi­tions I have for the program–enabling it to mod­el mul­ti­ple sec­tors and mul­ti­ple economies at once.

If we can raise anoth­er $30,000 or so, we can also address one of the main com­plaints that I hear about Min­sky: to quote my good friend Tom Fer­gu­son, INET’s Research Direc­tor, from our din­ner togeth­er in Lon­don last month, “Why is Min­sky so hard to use?”.

As I told Tom over din­ner, part of the rea­son is that econ­o­mists have almost no expe­ri­ence with sys­tem dynam­ics soft­ware.

This was com­i­cal­ly illus­trat­ed by Noah Smith’s ques­tion to me “Tell me what the ‘wires’ do in the tool?” on Twit­ter about two years ago, when Min­sky’s devel­op­ment had just begun: like most econ­o­mists, Noah had appar­ent­ly nev­er seen sys­tem dynam­ics soft­ware before, and thought I had invent­ed the idea of using “wires” to lay­out equa­tions visu­al­ly (and he thought I’d made a log­i­cal error too, though I must say Noah seems to have cot­toned on to dynam­ic mod­el­ling since). Engi­neers find using Min­sky easy, because the flow­chart metaphor for equa­tion design has dom­i­nat­ed engi­neer­ing now for over 3 decades.

But the oth­er part of why Min­sky is “so hard to use” is that “so lit­tle mon­ey has gone into its devel­op­ment”. There’s just over one man-year’s worth of cod­ing in Min­sky, so in some ways it’s amaz­ing that it does any­thing use­ful at all–let alone enough to win Source­Forge’s “Pro­gram of the Month” award back in Jan­u­ary 2014. My orig­i­nal appli­ca­tion to INET asked for US$250,000; I received US$128,000; with that mon­ey expend­ed, Min­sky just worked but had lots of bugs and many incom­plete fea­tures. With the addi­tion­al $100,000 raised from the pub­lic since, it now works rea­son­ably well, still has some bugs, but needs lots more fea­tures to be com­plete. Addi­tion­al mon­ey from you now will help add some desir­able “ease of use” features–such as direct entry of equa­tions, lay­ing out causal loops pri­or to build­ing a mod­el, bet­ter graph­ics, export­ing mod­els in vec­tor graph­ics for­mats for doc­u­men­ta­tion, etc. An absolute bomb of funding–if some hedge fund out there that uses my eco­nom­ic ideas (you know who you are) decides to give some­thing back–will enable data import­ing, fit­ting mod­els to data using non­lin­ear para­me­ter esti­ma­tion tech­niques, and so on.

So please, if you’ve read this far, and you sup­port what Min­sky is try­ing to achieve–giving eco­nom­ics the capac­i­ty to build real­is­tic mon­e­tary, dynam­ic mod­els of the economy–then go that lit­tle bit fur­ther and make a dona­tion via this wid­get, or via direct cred­it into Rus­sel­l’s accounts using the info in my invoice below: