Robert McDowell, writing at GitHub yesterday:

Some projects developers and I asked months and even years to get a very stable version, which includes sometimes millions line of code. As you can see, for various valid and logical reason, these projects won’t be developed again and again for the only reason that the development “zeitgeist” is changing every 10 years. It’s totally absurd and even dangerous for the freedom of Internet.

How exactly deprecating a proprietary system is dangerous for the freedom of the Internet is beyond my ability to comprehend.

Robert goes on:

I recall to develop a Perl application 18 years ago to automate invoices of a payment system for a company, during this time I had to change a very few function names and logic, but today it’s still working flawlessly. The web SHOULD be like that, don’t waste developers to reprogram everything decade after decade, it does not make sense at all, especially when the major tech used is not better and even worse than the previous.

In what other realm besides software is there this persistent expectation that something written decades earlier should continue to work in perpetuity with no ongoing maintenance? The Golden Gate Bridge, if it were not maintained regularly, would collapse into San Francisco Bay in a frighteningly short amount of time. This is the broken world in which we live.

If software doesn’t require ongoing maintenance, that’s a strong indication that it’s not being used, or that it’s not even usable, or that it’s retirement is drawing nigh, because something better is planned to take its place.

How important to its authors is the code that constituted Instagram when it launched in 2010? If we ran a git diff from the launch commit against the current tip of master, what would we see? Code is only valuable when it is serving its purpose in its time.

This fascination with relics is quaint. Civilization needs its museums and cemeteries but they needn’t be functional; they need only serve as reminders of the past and pointers to the future. If these people want to open source flash on a slow-to-boot VM so they can serve it up for a trickle of visitors who want to see how bad things were in the past, they’re welcome to try to do that; but they mustn’t kid themselves that this is anything other than a museum that will become passé as a new generation of developers grow up with vastly superior tools. Those people won’t care less about flash or the things that were built with it. The only thing I recall using that was powered by flash was E-junkie and it switched around 2014 after years of people begging the company to switch.

Which brings me to this part about “major tech” being “even worse than the previous.” It’s laughable. Nobody with a straight face can say that obsolete Flash on the web is actually better than Angular, React, Vue, and D3. It’s not even in the same ballpark.