Open-source software: Untangling code | Hacker News Reader

Open-source software: Untangling code

economist.com

18 pointsdavidw10 comments

I haven't read the book, but this article doesn't really say anything that any clear-minded open source advocate didn't already know. Outside of certain dogmatic circles, it's become pretty clear that the future of open source development is in the corporate world for many projects. That's not necessarily a bad thing, either.

I think the novelty was in the actual research they carried out that seems to confirm what everyone's noses had been telling them.

Here's a theoretical argument that seems unavoidable to me, though I haven't seen any evidence for it: many open source developers of a project consult in its area. It makes sense that that project is hard to use.

The great appeal of software, for me, is that it can encapsulate knowledge and expertise; a kind of abstraction layer. A magic talisman of wisdom and talent. This can include guiding the user in how to approach the task, implicitly providing concepts with which to reason about it. Like one of those old-fashioned how-to books, it can aim to be teacher, mentor, friend.

The "problem" is that if you give it away, you can't make any money on consulting, because (ideally) it itself is the best possible consultant. So you have to charge for it. But it doesn't seem quite so noble...

I'm reminded of a recent statement from one of the github guys to the effect that by running a profitable business using proprietary software, they're able to contribute more to open source.

I know it's hard to write pithy headlines, but describing open source as "programs written by volunteers" is misleading. A lot of open source software is actually written by people who get paid to do so, or who get paid at least partially because they do (or did) so.

         The Software Iceberg

                 /\
                /  \ 
               /    \ 
              / open \
             / source \
  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
           /            \
          / proprietary  \
         /  & for sale    \
        /                  \
       ----------------------
      /                      \
     /       proprietary      \
    /    & written in-house    \
   /                            \
  --------------------------------

Funny, I have the pyramid exactly the other way: open source should be the infrastructure, the base the 'crown jewels' at the tip of the pyramid are built on top of. Especially for 'lean startup' types of deals, where the lines of code for your startup will be several orders of magnitude less than what it's based on: Linux, postgres/mysql/nosql, tcl/perl/python/ruby/php/erlang etc....

I was referring to volume of code written, not function.

You bring up an interesting (and IMO valid) observation.

I guess this picture was only worth 999 words.

Volume of what? I think your pyramid is correct in terms of revenue, although maybe it's not so much a pyramid as a fairly stark line.

Volume of code, lines written/bugs created ;)

In house software typically has 0 revenue.

Reply on news.ycombinator.com