Such a conversion allows old code to run, albeit with an ever-increasing startup delay, as newer browsers convert old code to the new syntax. To avoid that delay, the site only needs to take the output of the code-transform and paste it in!
The effort has immediate benefits for those businesses interested in the results. The rest can sleep tight: their code will continue to run.
If we provide backward code-transformation also, then elder browsers can also run ANY new code!
Migration scripts should be created by those that make changes to a language. Today they don't, which is in itself a fundamental omission! It should be am obvious part of their job to provide them, as their job isn't really done without them. The onus of making it work should be on them.
With this system Any site will be able to run in Any browser, but new code will run best on the newest browsers. This way we reap the benefit of an up-to-date and productive development environment, where today we suffer, supposedly because of yesterday.
As an example, I have seen Patrick Michaud tackle the migration problem in PmWiki. It included forward migration scripts. Whenever syntax changes were made, a migration script was added to transform pages to the new syntax. As far as I know, ALL migrations have worked flawlessly.
In other words, we don't tackle the migration problem, we just drag it along. We are incompetent! And why is that? Because technically incompetent people feel they must decide for us. Because they are incompetent, fear rules them. They are obnoxiously conservative, and we suffer the consequence of bad leadership.
But the competent don't need to play by the same rules. They can
(and must) change them. They are the path forward. It is about time to leave the past behind, and pursue the leanest meanest, no, eternal functionality.
Even if the browser doesn't hook into this system, coders could. So language updaters should take it upon them to provide migration scripts. Once they exist, browsers may take advantage of them.