Too big, too pricey, too late

Too big, too pricey, too late

NASA astronaut repairing the Hubble in December 1993 • NASA (Public Domain)

Originally published 7 February 1994

The Hub­ble is out of trou­ble, NASA assures us.

A shut­tle crew has suc­cess­ful­ly repaired the flawed space tele­scope. Just look at those wide­ly pub­lished pic­tures of the galaxy M100. Before the repair, a mud­dy blur. After, crisp as a cracker.

Now sci­en­tists can “start solv­ing some of astron­o­my’s great­est mys­ter­ies,” gush­es Time mag­a­zine: “How old is the uni­verse? Do giant black holes lurk at the cores of galax­ies? How did the galax­ies get formed? Are there plan­ets cir­cling oth­er stars?”

It’s the NASA pub­lic­i­ty engine at full throt­tle. As if the Hub­ble is the great­est thing for astron­o­my since Galileo turned his home­made tele­scope to the sky.

Don’t get me wrong. I think the space tele­scope is ter­rif­ic, and the repair impres­sive. But the orig­i­nal bun­gle and the bal­ly­hooed fix are symp­to­matic of what is wrong with this kind of sci­ence. It is too big, too expen­sive, and too late.

Too big. The space tele­scope uses brute force — the megath­rust of NASA rock­ets — to accom­plish what is essen­tial­ly a mat­ter of finesse: mak­ing del­i­cate images of dis­tant objects.

Too expen­sive. Per­haps $3 big bil­lion, repaired.

Too late. The project was decades in plan­ning and exe­cu­tion, and in the mean­time tech­nol­o­gy passed it by.

What would those bil­lions have accom­plished had they been doled out in more mod­est por­tions to a broad­er seg­ment of the astro­nom­i­cal com­mu­ni­ty? Bet­ter light detec­tors. Sophis­ti­cat­ed dig­i­tal tech­niques for improv­ing images. Com­put­er-con­trolled cor­rec­tive optics to com­pen­sate for the blur­ring effects of the atmos­phere. Liq­uid mir­rors (a rotat­ing dish of mer­cury will adopt pre­cise­ly the par­a­bol­ic shape required of an astro­nom­i­cal mirror).

As the Hub­ble’s costs rock­et­ed, some­thing else got cheap­er down on Earth. Com­put­ing pow­er. What used to be accom­plished in astron­o­my with tons of glass and steel can now be done by calculation.

An exam­ple: Twen­ty years ago, the mir­rors of tele­scopes need­ed to be mas­sive to ensure the con­stan­cy of their shape. But the heav­ier they got, the more grav­i­ty caused them to warp and sag as they moved around. That was one advan­tage of plac­ing a tele­scope in orbit: no appar­ent grav­i­ty. How­ev­er, com­put­ers can now sense mir­ror dis­tor­tions and cor­rect them as they occur. Cheap­ly. On the ground.

The same for blur­ring effects of the atmosphere.

In oth­er words, much of the rea­son for going into orbit has van­ished dur­ing the time it took to get there.

If the Hub­ble images are bet­ter than present ground-based images, it is a mar­gin­al and tem­po­rary advan­tage, not a rev­o­lu­tion. If ground-based astronomers had spent the same kind of mon­ey they might already have sur­passed the image qual­i­ty of the repaired Hubble.

So for NASA or Time mag­a­zine to say that astronomers can now start solv­ing ques­tions such as “How old is the uni­verse?” is pure bunk. Astronomers start­ed answer­ing that ques­tion a long time ago, as well as the oth­er ques­tions list­ed by Time. The Hub­ble will undoubt­ed­ly con­tribute to our knowl­edge, but I will be very sur­prised if it dra­mat­i­cal­ly mod­i­fies our under­stand­ing the universe.

Then why does NASA make such extrav­a­gant claims? Well, if you had spent upwards of $2 bil­lion of the tax­pay­ers’ mon­ey for an inex­cus­ably flawed tele­scope, and maybe anoth­er $1.2 bil­lion to fix it, you’d be push­ing all the but­tons too. Yes, of course NASA is proud that the tele­scope has been repaired, and we should be proud with them; it is a spec­tac­u­lar achieve­ment. But we should keep in mind that the era of mul­ti-bil­lion-dol­lar sci­ence exper­i­ments is com­ing to an end.

It used to be that there were two branch­es of sci­ence: the­o­ret­i­cal and exper­i­men­tal. The­o­ries were affirmed and refut­ed by exper­i­men­tal obser­va­tions, and exper­i­ments often sug­gest­ed new the­o­ries. As sci­en­tists probed into the heart of atoms and out­ward to the edge of time, the machines nec­es­sary to do the exper­i­ments became ever more expen­sive. The recent­ly can­celled super­con­duct­ing super­col­lid­er, pro­ject­ed at $11 bil­lion, is a case in point.

Now there is a third branch of sci­ence: com­pu­ta­tion­al. Fast, cheap super­com­put­ers make it pos­si­ble to do a kind of sci­ence that is not quite the­o­ret­i­cal and not quite exper­i­men­tal. Com­plex sys­tems — the Big Bang, galax­ies, bio­log­i­cal mol­e­cules, the human brain — can be rep­re­sent­ed in com­put­ers as dig­i­tal data, and the com­put­er cal­cu­lates how the sys­tem behaves. The results are com­pared with the real world.

This rev­o­lu­tion of com­put­ing pow­er will inevitably lead to new dig­i­tal the­o­ries of real­i­ty, tru­ly rev­o­lu­tion­ary ideas about what makes the world tick (chaos, frac­tals, and cel­lu­lar automa­ta are hints of things to come).

The­o­ries of the future will be devel­oped and test­ed with the one kind of tech­nol­o­gy that is get­ting cheap­er with each pass­ing year: com­put­ers. And the finesse of com­put­er cal­cu­la­tion, rather than brute force, will car­ry more of the bur­den of exper­i­men­tal research.

Already, com­pu­ta­tion is trans­form­ing astron­o­my, as it is chang­ing every oth­er field of research from neu­ro­bi­ol­o­gy to pro­tein chem­istry. The Hub­ble space tele­scope is not so much the lead­ing edge of a rev­o­lu­tion, as the trail­ing edge of yes­ter­day’s science.


The Hub­ble Space Tele­scope is still oper­a­tional as of 2021, after five ser­vic­ing mis­sions by NASA astro­nauts. Its mul­ti-bil­lion dol­lar suc­ces­sor, the James Webb Space Tele­scope, is sched­uled for launch in late 2021. ‑Ed.

Share this Musing: