When the Web came about - I'm old enough to clearly remember the time before the web with Fidonet and Mailbox networks - I didn't really believe it would take off. Mostly because in Germany where I live using the phone would cost 20 pennies for 8 minutes back in the day in West Germany.
Like quite a few experts I completely underestimated how important colorful pictures to click on were for the vast majority of regular people. I only really got into the Web when it already had gotten critical mass, roughly around 1999.
The Web - or at least its technology - has clearly won as a platform. Human readable (more or less), cross-platform and open standards + open source. Nice. All the good stuff.
However the Webs shortcomings are still the exact same as they were 30 years ago, some related to the internet in general:
unfitting structure of DNS (needs a redo), no independant SPOT registry with hard crypto authenticity (that would actually be a rare truely useful usecase for blockchain), always online/no true offline capability, encryption and signage only as a flaky afterthought and not a hardwired fundamental feature/requirement, non-deterministic / non-existant control of visuals, and a few other details that are a huge PITA and likely the cause for 80%+ of webtraffic today.
I've been doing professional web development for 24 years now and can attest that there are fundamental things that really could use a complete overhaul.
We need:
- independant DNS with crypto auth/auth and a signed registry (waaay overdue ... as I said, some blockchain thing might even be useful here)
- async and offline and perhaps hard standardized versioning built in as a main feature (OMG sooo needed)
- large cross-platform standard library of fonts & icons (it's not that we don't have them these days and client storage space really isn't an issue anymore)
- cleanup and polishing of language and culture/country zones (that is already pretty good I admit, but it could use a little polishing)
HTML & CSS need a complete cleanup
- absolute sizes (and nothing else! ... scale yourself however you like it ... Pixels and print standards from the steam age are a thing of the past, let's make it that way!)
- no magic numbers (h1 - h6 ... WTF???)
- sane hard document structure requirements with hard errors and no further relay for broken documents (here a versioning standard would come in handy)
- client side turing logic / visual manipulation built in (this is a tricky one because it will be abused but perhaps some basic DOM manipulation for limited interaction would make sense)
That's just from the top of my head. The Web has aged and it has aged pretty well, but the amount of hacks we have to do to make stuff happen and fix the broken things is quite astonishing. ... Given, that's the reason I earn obscene amounts of money but technically we could do all this easier.
By and large Google is just a better Altavista. If their math experts can't keep up with the spammers or if Google management decides it has to do my thinking for me and does a shitty job at it I'll just move on to another search engine. I'm sure there are folks gladly ready to take over if Google drops the ball.
But a Web redo as laid out above would fundamentally fix the things that always were broken. And that would be a huge improvement and likely have some aspects of search engines become superfluos. Tough for Google and their 'little' ad business, but overall a good thing.
2 cents from an senior webdev.