25 years ago: Sir Tim Berners-Lee builds world's first website There are just under a billion web domains registered in the world today, and over four billion webpages, by some estimates. We've come a long way: it all started to come together just 25 years ago in a small office at the European Organization for Nuclear Research (CERN). On December 20, 1990, a Fellow at CERN, Tim Berners- …
Yes. And URI scheme names are canonically in lowercase, so it should be "the http: in URLs", not "the HTTP:" as the article has it.
(And, arguably, it should be "in URIs", though HTTP URIs are always URLs,1 and there's some sentiment in the W3C and IETF URI working groups that we should deprecate the URI abbreviation and simply use the term "URL", as the URL/URN distinction is not widely recognized. And many careful writers would prefer that "http:" in the phrase above be enclosed in quotation marks, since we're talking about the literal string per se and not the URI scheme it names.)
You might think a tech site would go to a bit of trouble to get the technical details correct. Unless you're a regular reader, of course.
'tis a good day for snark indeed.
1More precisely, they're always syntactically URLs, even though they're not always semantically URLs. That is, some times they do not in fact "locate" a resource, but are used purely as names. It's a largely-pointless distinction only a computer-scientist-philosopher could love. I like it a lot.
I remember when it was the "World Wide Wait" because transfer speed was so slow, watching a page load was pretty close to watching paint dry.
I remember seeing a PC struggling with NCSA Mosaic in the uni labs and thinking "this'll never fly"
As usual, I've kept my 100% score of being wrong on major things.
and ISDN line would put you in debt for life and a 10meg hard drive was a lot less.
I'm not sure where the phrase 'never under estimate the bandwidth of a lorry load of tapes' came from but the internet didnt seem viable,
Having said that I had a 2.4Gbit fibre optic 't switch' for $5 rejected cos the company decide it didnt do that sort of thing any more and now they pay a fortune for slower heaps of shit.
Andrew S Tanenbaum in his book 'Computer Networks'.
Specifically the final paragraph of 2.2.1, "Magnetic Media". Page 57 in the second edition (1989).
The book's older than the WWW and still relevant. Of course things like the Nyquist Limit don't go out of fashion.
… is that the whole thing is open. HTTP, HTML and all of the protocols that underlie the web are free for everyone to use, to analyse and to be involved in. Even the networking protocols which carry all of the traffic.
It’s worth noting while modern IT companies make truck loads of cash charging fees and suing everybody else how much we have all benefited from a free and open Internet.
"modern IT companies" also gave us most of those protocols used in the "free and open Internet". Yes, some - such as HTTP - came from research institutions and academia; but a great deal of it was the result of work done at for-profit organizations, and paid for by a combination of government grants, investors, and profits from commercial IT products.
It's the result of a huge and diverse community of contributors. (And a great many efforts went nowhere or flamed out after a brief period of use, too.) There's no single, simple economic system that produced the modern Internet.
Happy Birthday to you
Happy Birthday to you
Happy Birthday Dear http://www.whatever
Happeeee Birthday toooooo you.
Just another example of my long held belief, that although IT stuff is built on the shoulders of all our predecessors. Real advances are made by individuals or very very small teams.
@Altogether Now
"Real advances are made by individuals or very very small teams."
As are all things useful.
My long held belief over the span of 50 years in IT.
As a project manager my boss contributing more headcount was a bad thing, slowing the project down rather than accelerating progress.
Oddly enough, with the rise of single-page javascript web applications the number of "pages" per website is dropping over time... although the amount of information shown via them continues to rise.
Indeed the biggest barrier of single page web apps is the fact that a full page reload makes it easy to cycle even more adverts into our eyes. Which means reloading loads of mostly-unchanging content each page reload and worse, reinitialising a load of javascript.
For example, this website could be a single page webapp (online article viewer) with most top-links being filters, and the stuff on the right being direct links to articles (again, presented by the webapp). The only other pages would be the mostly static stuff linked in the footer.
with the rise of single-page javascript web applications the number of "pages" per website is dropping over time
Yes, one problem is that the metric for "web page" is not well-defined. Static pages are easy enough to count (conceptually), but as you say they're increasingly rare. Do you count distinct URLs that produce distinct resources, as measured by their output? Even if you screen out obvious "error pages", though, there are link-farming sites that generate pages for any URL; you have to screen that crap out too. And then you have to deal with the dynamic content (like ads and "personalized" content) that differs between retrievals of the same resource, as named by a given URL.
So really all these "X pages on the web" figures are just pulled out of the air.
In any case, what does it matter? Milton is said to have read every book available to him in English, French, Latin, and Hebrew (I think that's the list). No one today can possibly read a sizable fraction of all the web pages available in a single language. There's more out there than any one of us will ever see. Exactly how much is irrelevant for most purposes.
Now, if you're, say, trying to archive the web, it's of some interest (as are all those issues above with dynamic content). But for ordinary punters? Just make up a big number. I say there are 46827356492 web pages, as of right ... now.
As a student I was working just down the corridor from him in December that year. I remember one of the computer students going on about how a new service to rival ftp and gopher was being created. I ignored her and concentrated on my experiment and the beers from restaurant 1 (is *that* barman still there?). Another world shattering event that I managed to be near and completely ignore.
He may have wanted only text - but the first thing that went on it was pornography. The first site that I actually saw was over the shoulder of a high energy physicist a year later and it was showing a lady doing something unmentionable with a beer glass - prompting the comment from a fellow female student: 'at least she didn't try that with a glass with a handle on it'.
Put up some data
There are many ways of doing this. The web needs both raw data -- fresh hypertext or old plain text files, or smart servers giving views of existing databases. See more details , etiquette.
Suggest someone else does
Maybe you know a system which it would be neat to have on the web. How about suggesting to the person involved that they put up a W3 server?
________
If only I could think of a neat idea. In plain text obviously. Then this W3 thing might take off.
LaTeX? At the time, formatting a few ages of LaTeX into DVI took several minutes on a VAX (never mind printing it). Presumable Sir Tim wanted something simpler to make the system interactive. I am also guessing he wanted to be compatible with SGML, which at the time was touted as the final solution to documentation problems. HTML syntax is based on it. The earlier HTML specifications actually formally represented the language as an application of SGML, which is (or was) a kind of general toolkit for specifying document formats.
The earlier HTML specifications actually formally represented the language as an application of SGML
HTML up through version 4 is an SGML application. It has two serializations: a straight SGML one, and an XML one. And since XML is an SGML application, both of those are SGML applications.
HTML 5 has its own serialization, which is not an SGML application AFAIK, and an XML serialization, which perforce is an SGML application.
There were a number of good reasons to make HTML a (quite simple) SGML application. SGML is an international standard, and was back when HTML was invented. Numerous SGML parsers existed, for a wide range of platforms. It has a long history, evolving as it did from IBM's GML, which in turn was inspired by DEC RUNOFF (which also inspired UNIX's roff). It's easy to use and forgiving, though unfortunately that led to a vast corpus of invalid and poorly-written HTML content. It has mechanisms for precise specification and machine validation, but casual practitioners don't have to know about them. It's extensible and flexible.
As someone else pointed out above, LaTeX, despite its advantages for text and mathematical layout, is much more resource-intensive and slower; it also was overkill on the relatively low-resolution graphical displays of the time, to say nothing of text-mode browsers. Decent LaTeX output sometimes requires changes to the input - no one wants to throw together a quick web page and get "overfull hbox" errors.
As it is, we can be grateful that TBL didn't use Display Postscript.
I'll also add that for mathematics on web pages we now have MathML and TeX mathematical typesetting via Javascript, so that particular problem is pretty well settled. It seems to me that most sites these days that want to typeset mathematics just use MathJax, which lets you use TeX $...$ notation or MathML or some other things. That's convenient for page authors and for people posting UGC (comments and such).
"In the long term, when there is a really large mass of data out there, with deep interconnections, then there is some really exciting work to be done on automatic algorithms to make multi-level searches. "
Nomination for understatement of the last few decades.
The rest of the CERN site is interesting reading as well, brings back memories...
Tim knew the potential on his hands. A few notes from his prototype website (http://info.cern.ch/hypertext/WWW/Bugs.html):
--
"Things to be done"
Form processing
If you can edit hypertext, you edit a hypertext form and return it. To be able to submit a form back to the server would allow special search patterns, administrative processing, electronic voting, ...
Search engines
Now the web of data and indexes exists, some really smart intelligent algorithms ("knowbots?") could run on it. Recursive index and link tracing, Just think...
--
"Just think" indeed Tim! I can't help but read that and feel a proud sense of nostalgia.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
