Posted by Gavin Doughtie
Sat, 17 Dec 2005 05:36:00 GMT
I’ll be 43 this Sunday. I write code every day. C++ Linux graphics code in the daytime, Ruby on Rails AJAX stuff at night.
Here’s what I think. Guys my age have mortgages and children. We need to earn at the higher end of our capacity. For some guys that means moving into management (for organizations that value management over engineering). For some it means developing ever more arcane specializations.
If you’re a twentysomething IT developer, you aren’t even going to see us specialized guys. We’re busy with problems that involve domain knowledge you’ve never needed to learn, fiddly weird hardware, languages that are too low-level, or too high-level to be part of mainstream curricula and entry-level jobs.
I ran a great number of developer interviews at one of the start-ups I worked at a few years ago. Almost all the candidates, young and old, were duds—but sometimes some graybeard engineer would show up and just floor us with his knowledge and understanding. Inevitably, he’d get a better offer, and vanish again into the rarified realms from which he descended.
I don’t think age is the issue. Value is. If you keep learning and remain valuable, there will be opportunity.
Posted in Software | 2 comments | no trackbacks
Posted by Gavin Doughtie
Fri, 28 Oct 2005 21:32:00 GMT
I remember I met a mainframe programmer once who could compile IBM assembly code in his head.
Do you think he really misses doing that now?
People are creative, resourceful, pleasure-loving, inventive and funny.
So, why do we ennoble jobs that turn these wonderful creatures into interchangeable machines? Why do we fight to make assembly-line jobs high-paying?
We should be fighting to make it unprofitable to employ human beings in any endevour that is unpleasant, dangerous, demeaning or dull. Ancient Greek steam engines were never adopted because slaves were too cheap. Let’s go the other way, now that we live here in the future. Let’s push to make automation so cheap that even slaves are more expensive. (That way, you see, there’s no reason to have slaves).
Posted in Creativity, Software | 2 comments | no trackbacks
Posted by Gavin Doughtie
Sat, 08 Oct 2005 20:39:00 GMT
So the demos I've been doing at Web 2.0 of
xdraw,
my shared vector-drawing application, have been using the fairly brain-dead "web page polls the server continuously for changes" approach to AJAX chat. Naturally, this is pretty bad for the health of my server (50% CPU consumption anyone?).
The better approach is to have the server hold open the connection and jam data down the pipe whenever something changes. This is the system used by
jotspot and the
HTTP.push
library.
Stay tuned as I turn it on.
Posted in Web 2.0, Software, Ruby, AJAX, Javascript | no comments | no trackbacks
Posted by Gavin Doughtie
Fri, 07 Oct 2005 19:15:00 GMT
Just watched a demo from Mental Images, where a 3D visualization was rendered in real time on a server, and the resulting images streamed back to the web browser.
The images themselves were transmitted with lowres/hires pairs, so that when you stopped moving the mouse, you’d see a more bandwidth-intensive and pretty picture.
Now, this frees up the content owner to render from a scene that could possibly have gigabytes of data, while only sending small image files to the client.
But, really folks, it’s a nasty hack to make up for lack of client side 3D apis.
That got me to thinking, why don’t we have this? Ten years ago this summer, I remember sitting with the VRML dudes planning our new 3D future.
Didn’t happen. Here’s my theory:
At the time, the client was considered the “property” of client-side software companies. Thus, despite efforts at building consortia, the 3D plugins were developed and delivered as proprietary software which (partially) implemented a standard specification. Each software vendor wanted a proprietary advantage, and nobody opened their source. The business was never conceived of as something built atop an open, interoperable framework. It was culturally unsellable in a shrinkwrap software world.
It wasn’t just the wacky 3D browser world, it was more serious and potentially more useful technologies such as the now lowly Java applet.
Here’s what I think is coming. The business has moved from the browser to the server over the last decade, and all that money wants customers to have a great end-user experience. More importantly, a consistent end-user experience no matter what OS choice the user makes. So, we’ll fund Firefox and KHTML, sometimes with cash but more often with intellectual effort, to give us a better outlet for our businesses.
Posted in Web 2.0, Software | no comments | no trackbacks
