Let's Innovate

Back from spend­ing three days at an invitation-only event hosted by the National Cen­ter For Super­com­put­ing Appli­ca­tions (NCSA) in Cham­paign, Illi­nois. NCSA is one of the old­est of the National Sci­ence Foundation’s Super­com­put­ing cen­ters, there being only three left of the orig­i­nal five (some peo­ple would say only two). For any­one too young to remem­ber it is the birth­place of Netscape, and thus in a prac­ti­cal sense, of the World Wide Web. It will soon be the site of Blue Waters; the first open sci­ence super­com­puter to achieve sus­tained 1 Petaflop per­for­mance on real sci­en­tific and engi­neer­ing appli­ca­tions. The meet­ing I attended was focused on encour­ag­ing com­pa­nies (par­tic­u­larly man­u­fac­tur­ing com­pa­nies with R&D needs) to take advan­tage of NCSA and its com­pu­ta­tional resources to shorten time to mar­ket by replac­ing tra­di­tional phys­i­cal test­ing with com­puter mod­el­ing and sim­u­la­tion. One of the speak­ers, Tom Lange (head of the mod­el­ing and sim­u­la­tion depart­ment at Proc­tor and Gam­ble) was par­tic­u­larly blunt: “For every dol­lar you give me to spend on mod­el­ing and sim­u­la­tion, I can save five dol­lars of tra­di­tional test­ing.” Equally impor­tant these days, of course, is the sig­nif­i­cantly reduced time-to-market that mod­el­ing and sim­u­la­tion enables for com­pa­nies want­ing to get new inno­v­a­tive prod­ucts out the door.

It was a fas­ci­nat­ing meet­ing, and I’ll have a lot more to say about it in the com­ing weeks as I have time, but first I’d like to focus on a theme I heard expressed by sev­eral speak­ers, and which I think is impor­tant. We’ve all heard about the impor­tance of increas­ing the num­bers of grad­u­ates in sci­ence and engi­neer­ing. The National Busi­ness Round­table esti­mates that the United States should be grad­u­at­ing 400,000 under­grad­u­ates per year with tech­ni­cal degrees. The cur­rent num­ber hov­ers around 265,000 and a severe short­age is loom­ing. Com­pound­ing this prob­lem is the fact that vir­tu­ally none of the BS and MS grad­u­ates in any of the engi­neer­ing sci­ences are com­ing out of school these days with any knowl­edge or skills in mod­el­ing and sim­u­la­tion. The num­ber grad­u­at­ing with the abil­ity to use super­com­put­ing mod­el­ing and sim­u­la­tion is essen­tially zero. Speaker after speaker noted the fact that sim­u­la­tion is some­thing that PhD grad­u­ates are some­times famil­iar with, but for the most part they aren’t hir­ing PhD’s. How­ever, when schools are asked about this gap, the answer is that there sim­ply isn’t room in the cur­ricu­lum to start offer­ing these courses at the under­grad­u­ate or master’s level. While this is obvi­ously cor­rect given the cur­rent cur­ric­ula, it almost cer­tainly reflects a con­flu­ence of mis­placed “aca­d­e­mic rigor” and some­thing like: “No under­grad edu­ca­tion is com­plete if they don’t have to take the same courses I did.” But, the fact of the mat­ter is that what con­sti­tutes a rig­or­ous, com­plete edu­ca­tion has changed, and has changed at all lev­els. No one ques­tions allow­ing cal­cu­la­tors into high school math classes and math tests today, and any­one sug­gest­ing a class in slide rule use would be laughed out of a cur­ricu­lum meet­ing. At the under­grad­u­ate engi­neer­ing level, I believe it is time to exam­ine the amount of time that engi­neer­ing and sci­ence cur­ric­ula spend study­ing cal­cu­lus. Math­e­mat­ica, Math­cad, Maple and sim­i­lar pieces of soft­ware should be tools for not just enhanc­ing stu­dent appre­ci­a­tion of func­tional rela­tion­ships, they should be allowed to replace the grind­ing hours spent learn­ing ana­lytic inte­gra­tion or solv­ing dif­fer­en­tial equa­tions in closed form. The real­ity is that the num­ber of real world prob­lems that can be mod­eled by a dif­fer­en­tial equa­tion that can be solved in closed form is infin­i­tes­i­mal – and the few that can rep­re­sent very unin­ter­est­ing problems.

Then, let’s take this process one step fur­ther – let’s push both the intro­duc­tory sci­ence and engi­neer­ing and the intro­duc­tory mod­el­ing and sim­u­la­tion down into char­ter high schools and advanced place­ment (AP) classes in our reg­u­lar schools.  I can think of no bet­ter way to stim­u­late inter­est in sci­ence and engi­neer­ing than to expose our best and bright­est stu­dents (who already are won­der­fully tech­ni­cally– and computer-literate) to the pos­si­bil­i­ties that open up when given the chance to explore a vir­tual world of devices they them­selves build using workstation-level sim­u­la­tion tools.

Many of you are rolling your eyes right now – “Is he crazy?  Does he expect high-school kids to be able to use Nas­tran?”  In a word, yes.  It can be done, has been done, and , and needs to be done more often in more places. Not every high-school kid is going to want to play with NASTRAN (or PATRAN, ANYSIS and the rest of the mod­el­ing tools that exist today), but the AP stu­dents who grad­u­ate from high school with the equiv­a­lent of two years of col­lege level cal­cu­lus cer­tainly can, and as part of expos­ing them to the excite­ment of what com­put­ers can do, they deserve a chance.  What’s more inno­v­a­tive pro­grams like “Project Lead the Way” are show­ing just what high school and mid­dle school stu­dents can do when chal­lenged with sci­ence and engi­neer­ing when it is pre­sented as the excit­ing adven­ture it can be.

All this would be strictly smoke blow­ing, hand wav­ing, and pie-in-the-sky wish­ful think­ing, except for a con­flu­ence of three other events.  First, as part of the and the “State Fis­cal Sta­bi­liza­tion Fund”, the cur­rent admin­is­tra­tion is mak­ing $44 Bil­lion avail­able for revi­tal­iz­ing Amer­i­can edu­ca­tion.  Sec­ond, the new Sec­re­tary of Edu­ca­tion, Arne Dun­can has a rep­u­ta­tion for being will­ing to “think dif­fer­ent” – he has and is pro­mot­ing char­ter schools, has not hes­i­tated to close under per­form­ing schools in Chicago, and in Chicago has suc­ceeded in achiev­ing sig­nif­i­cant improve­ments.  Finally, direc­tor of the most suc­cess­ful char­ter school orga­ni­za­tion in the coun­try, the “Green Dot” schools, is being con­sulted about expand­ing the model to have national scope.

The time is ripe for change. All that is required is for some­one to put these three pieces together.

Tags: education, HPC, NCSA

You can leave a response, or trackback from your own site.