He never intended PHP to last longer than 6 months. He kept waiting for someone to make something better so he could shut it down... it just didn't happen. Also, he admits he's not a great programmer.
As a counter point doesn't every language "introduce problems" of some kind?
That said, the improvements in version 7 look impressive. I still prefer a language that's not quite as ugly though.
But a lot of the choices were short sighted, and a devotion to backwards compatibility has limited the ability to clean up.
PHP is now a super fast, relatively mature web programming language that carries a lot of its baggage.
He addresses your issues about "short sighted" choices and backwards compatibility.
> Please don't comment on whether someone read an article. "Did you even read the article? It mentions that" can be shortened to "The article mentions that."
I was trying to suggest it as a helpful way to understand Rasmus' perspective and a very complex history of development that has real logic to the choices he made that so many people have been critical of over the years. And that I think if someone heard him out would realize they would likely have made the same choices he did if they were in the same situation.
I just felt there wasn't a need for me to quote the video, but I suppose that was a little lazy of me.
Some folks were also compiling C and gluing it together with a CGI module, and a lot of others in the Microsoft world were using Frontpage and its brethren. But, the Microsoft stack only ran on Windows servers, and Windows web servers had a reputation (deserved, imo) for kinda sucking.
PHP's killer feature was mod_php. It was dead simple to integrate into any shared hosting environment, and so pretty much every shared host did. mod_php magically knew how to spit out html when reading html, and execute php when encountering php tags, so that gave us the html-mixed-with-php pattern that was wonderful at the time but considered bad practice now.
But also scope has changed a lot. At the time, people were mostly looking for hit counters, contact forms, and "guestbook" forms. For all those things, PHP excelled, so it gobbled up the market fast.
My first website back in `96 also had some compiled C via cgi. No database. I've said this before but using compiled Go with the web brought back memories of those good old days.
I spent about 4 years doing classic ASP before .Net arrived and had a hand in building a lot of big sites some of which still exist (and still use ASP): Most corporate jobs seemed to use the Microsoft stack then (or Java), where PHP was the what you did on the side because it didn't require server licensing.
Also safe_mode and open_basedir, which promised shared hosting providers that their users wouldn't hack the server and each other
Once you could get Apache with mod_php set up, PHP felt like something you could use to actually get stuff done.
There were other things like Zope and Roxen/Pike around back then, but I never used them. Ah and OpenACS/AOLServer which was TCL from memory.
PHP hit an optimum equilibrium between usability, features, and price. This only changed when virtualization changed the hosting game. Arguably PHP is still very close in terms of usability - deployment is fundamentally trivial - it's just lagging elsewhere.
I was primarily a Perl programmer at the time, responsible for writing CGI apps that were slow and unreadable to anyone I worked with who didn't know Perl.
I think vbasic was around but that required IIS and was an unmitigated disaster in every way. I sadly had to work in that a lot too.
I never messed with coldfusion but it seemed to get the job done for people who did.