Hacker News new | ask | show | jobs
by pdkl95 4361 days ago
First, even though I disagree with several of your points (GNU vision failure, your teleological view of innovation), I don't disagree with all (what's good for Google, #2 being strong currently).

Change (innovation) is hard to predict, and tends to happen in areas we least expect[1]. In particular, I fail to see why the people in the Free Software community would be any worse than those doing similar work in the corporate or government environment in security or other areas. In fact, there are several data points that suggest that are better in some areas.

The fallacy, I suspect, is the assumption that those supporting the "GNU vision" are a distinct group from the others, when in reality there is a lot of overlap between your three categories. A good example is IBM, who - upon remembering they are in the business of selling "solutions" and service contracts - opened up a lot their software, and participated a lot more in the FS and OSS communities.

Also, I'll mention that we should hope some version of your "GNU" future happens, at least where communications over the internet is concerned, as "proprietary" and "encryption" don't really work together if you ever want to trust it.

An alternative, vision of the future you may want to consider: most software is (and will be) fancy plumbing.

In each field, once you have the initial discoveries made (e.g. Knuth's tAoCP, the basic hydrological theories necessary to move water that we take for granted today) the rate of real discovery slows down considerably and becomes much more specialized. Innovation happens in both. New valves and fixtures are made, etc, but most of the time you just need to hire some plumber to setup your building, or you need to hire some software professional to keep your data backed up and your invoices generated properly.

To achieve this, just like how "standard parts" are far more common than proprietary parts in plumbing, having an free and open body of software to draw on makes the equivalent job in software far easier, too. I suspect this will be true with the internet and without.

None of this suggests the corporate or government players will go away - they will simply gain the benefits a free and open library as well.

[1] first paragraph: http://en.wikipedia.org/wiki/Connections_%28TV_series%29#Con...

1 comments

> Also, I'll mention that we should hope some version of your "GNU" future happens, at least where communications over the internet is concerned, as "proprietary" and "encryption" don't really work together if you ever want to trust it.

While this future would be advantageous for me, I am not sure I am onboard with everyone being responsible for their own infrastructure. Not only is it quite wasteful in terms of physical resources (and very sub optimal on energy consumed), but it requires that every individual community have several experts with a high degree of skill (for reasons previously discussed of which I see no actual rebuttals)

> To achieve this, just like how "standard parts" are far more common than proprietary parts in plumbing, having an free and open body of software to draw on makes the equivalent job in software far easier, too. I suspect this will be true with the internet and without.

This is not an impossible future, but people have talked about it over and over, and every time a new language or paradigm comes out that is supposedly the "key" to unlocking truly modular and reusable software. Heck, even Brad Cox thought he was doing it with Objective-C and look at how much code is duplicated between iPhone apps.

> responsible for their own infrastructure

So you'd suggest we go back to the Ma-Bell model where you couldn't plug anything into the phone network yourself? Or do you like being responsible for your own phone and modem? While it was a clever hack, the acoustic coupler should never have been necessary.

> but it requires that every individual community

No, it only requires enough communities have such experts such that a "herd immunity" is established.

> I see no actual rebuttals

Your decision to totally ignore my point about these groups having a high degree of overlap is noted.

The assumption that "expertise" only happens with "commercial" (or government) was disproved a long time ago.

> modular and reusable software

"standard parts" isn't really about modular programming - I mean stuff like: Apache, glibc, zlib, etc.

Really, I'm just sketching out what is done every day to meet the mundane needs of "business software". A typical business generally doesn't need someone to invent a new data storage technique - they just need someone to make the necessary tables in postgres/oracle/whatever and the necessary forms/report-generators.

These needs never go away, and are the kind of task that benefits greatly from having a collection of commonly-used tools to draw from.