Google Reader is dead. This was a product that many different people used for different purposes. I used it mainly as a way to subscribe and listen to my favorite podcasts. Others used it as a way to integrate disparate data feeds into a manageable and browsable list. It started in 2001 as JavaCollect.
What occurs to me at 2:26am PST, July 2nd, 2013 is that this doesn’t last forever. The approximately 10-15 year life span of a product, which seems like forever in our current eyes, it just a blip in the screen towards the things we should be building.
I have been an itinerant member for a long time of the San Francisco-based Long Now society, http://longnow.org. This is an organization dedicated on so many levels to the development of artifacts and culture that will endure with permanence over the the course of the next 10,000 years.
Software should not be excluded from this endeavor.
We have to recognize that the era of Moore’s Law, in its simplest form, is drawing to a close. There are physics limits to how small and fast a cpu can get. Obviously its not the end of innovation and change, but the areas where such changes are occurring over the next 10-30 years will be different from the last 10-30 years. What this means is that technological platforms are going to stabilize. Much as electricians of 2013 use the same formulas for electricity developed 100 years ago, similar things will happen within industries based around computation. When you fast forward to a future society used to electricity and computation as a basic part of normal life, the algorithms and physics we have developed in our times will be more or less used in similar ways going forward.
So here we see a widely-used product that has been around for 10-15 years, based on definition. There are applications and platforms are are in existence now that will still be in use 100+ years from now. Perhaps longer. Imagine debugging code that is 200-300 years old. That will happen.
If you think that its marginally important that your restful API is truly restful, try to imagine who might be using it in 100 years, much less 10k. There *will* be API’s living in some form today that are still in use in some aspect 10,0000 years from now.
I suppose part of this post is in reaction to Google Reader shutting down. For what I personally used it for, it worked well for many years. I’ve looked at any number of alternatives and none quite get it right. Seriously, just copy the Google Reader interface and you’ll be at 100% of what I need.
On the other hand, since its (now) shut down at 2:44am, 7/2/2013, I am realizing what I *wasn’t* missing. The frequent problems of playing an mp3 and not having it finish loading. Chrome completely locking up and needing to be process-killed. I’m on a Mac. This should not happen.
I look past that though, and I see a piece of software that was *always*, *ALWAYS* reliable in making sure new feeds were updated when they were available. Sure, their interface was always something to be tweaked but it was about 94.3% of the way right for what I used it for.
I imagined being able to use Google Reader till the day I died. I’d be an old, old curmudgeonly man stepping once or twice a week into my old Google Reader account to see what was going on with the other old curmudgeonly people I grew up with. I can recognize that from the perspective of my early 30’s, there are some things that just seem comfortable and I see how our ancestors just used the things that worked when they were young and got to the age that they didn’t care to experiment any more.
So this most certainly is not an argument against innovation. But it *is* a warning to those of you reading this in your bright and immortal early to mid 20’s that software is, and is not, forever.
Just to wrap this all up, Google Reader is dead. It might have lasted for decades, but the people *currently* running Google made the choice that it wasn’t worth maintaining. So now its dead. I exported my data, did you? But I really don’t have an equitable place to send it to.
So, think about the software you create. What you think is a weekend hack may very well become the core of the software you release for the next 20 years. (I’m looking at you, Brendan Eich.)
Try to plan to build multi-generational businesses and software, not something that’s throwaway that you hope will be bought out by someone in a couple of years.