Hacker News Reader

Comment by eljimmy | Hacker News Reader

Viewing the source of that webpage really takes you back. Plain old HTML. It's nostalgic beauty.

I recently fixed up an old 486 I purchased off eBay but it was bittersweet when I managed to get it connected to the 'net. Most websites were inaccessible due to the lack of support for today's encryption protocols, those that were had numerous JavaScript issues.

View on HN

I do appeciate some of it, but I don't miss this type of thing:

  <font size="5" color="#9400d3">

Though I understand why they are doing it in this case.

Definitely was a lot of work without CSS.

The <table> layouts are definitely not missed by me.

Were you just using the browser/OS it came with, I guess like Windows 3.1 or 95?

Given their well-publicized insistence on building for a ton of obscure arches, I'd expect you could run modern Debian on such a machine no problem, with a modern web browser. Might be a little slow, especially if you stick with the original disk, but should be perfectly usable.

> I'd expect you could run modern Debian on such a machine no problem...

Nope. Current builds of i386 Debian require a Pentium Pro or later -- I believe it's because they're compiled with the CMOV instruction, which wasn't present in the Pentium or earlier.

Ah, good point. Looks like Debian Jessie might have worked for it? In any case, distrowatch suggests a few others like Alpine and TinyCore that might have the proper support.

Yeah I had installed Windows 3.1 and tried IE and Netscape Navigator Gold.

Linux might be a whole other battle to get working but might be a fun project to attempt.

Try NetBSD with Lynx. OpenSSL will last several seconds, but it will work.

Except this bullshit:

    <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 2.0//EN">
    <meta http-equiv="Content-Type" content="text/html; charset=utf-8">

    <html>

Back in the day, you only needed:

    <html>

Nitpick 1: That DTD is for HTML 2.0, which was published in 1995. If that does not qualify as "back in the day", I don't know what does.

Nitpick 2: <meta> goes inside <html> (inside <head>, really).

Nitpick 3: The <meta> tag is only a band-aid for shitty webhosting where you cannot access the webserver config to make it send the correct Content-Type in the actual HTTP response headers. The modern <!DOCTYPE html> instead implies a default of UTF-8 which works well for most.

Nitpick nitpick: the html doctype doesn't imply UTF-8. Valid modern HTML documents must be encoded using UTF-8, but the standard also requires that the encoding be specified somehow.

> The Encoding standard requires use of the UTF-8 character encoding and requires use of the "utf-8" encoding label to identify it... If an HTML document does not start with a BOM, and its encoding is not explicitly given by Content-Type metadata, and the document is not an iframe srcdoc document, then the encoding must be specified using a meta element with a charset attribute or a meta element with an http-equiv attribute in the Encoding declaration state.

<https://html.spec.whatwg.org/multipage/semantics.html#charac...>

Don't you only need

    <!DOCTYPE html>
    <html>

these days? That's slightly worse but not terribly so IMO.

Oh I actually quoted the website's source. They have that DTD meta crap in there.

But I think you can just do <html> nowadays and it empirically just works. Seriously, screw the anti-DRY people that want me to put some !DOCTYPE or xmlns tags with some W3C links or some DTD nonsense inside ... I should only have to specify "html" exactly once, no more.

If I had designed the spec I would have just made it

    <html version="4.0">
    <html version="5.0">
    <html version="5.1">

Incredibly more readable, and memorizable. A markup language (literally), by virtue of being a markup language, should not be impossible to memorize. Making scary strings like "-//W3C///DTD" part of the spec is counterproductive.

You don't even need <html>. <https://html.spec.whatwg.org/multipage/semantics.html#the-ht...>

This is a valid HTML5 document:

    <!doctype html>
    <title>This is valid!</title>
    <p>Really, it's valid!

Paste it into the validator yourself if you don't believe me: <https://validator.w3.org/nu/#textarea>

That's SGML tag inference at work (theoretically at least, since browsers have HTML parsing rules hardcoded). SGML knows, by the DOCTYPE declaration, that the document must start with an "html" element, so it infers it if it isn't there. Next, by the content model declared for the "html" element (normally obtained via the ugly public identifier that sibling comments complain about), a "head" element is expected, so SGML infers it as well if it's declared omissible, and so on.

In the "old days", web pages were often just the bare content (no html, head, body containers, no DOCTYPE declaration). A few sites also featured just the body tag (and respective content) for setting the background attribute for the page background color.

E.g., this is the entire code of Netscape's first home page:

    <TITLE>Welcome to Mosaic Communications Corporation!</TITLE>
    <CENTER>
    <A HREF="MCOM/index2.html"><IMG SRC="MCOM/images/mcomwelcome1.gif" BORDER=1></A>
    <H3>
    <A HREF="MCOM/index2.html">Click on the Image or here to advance</A>
    </H3>
    </CENTER>

http://home.mcom.com/

I LOVE <center> and still use it sometimes because it's too goddamn complicated to center anything days.

"margin-left:auto;margin-right:auto;" what the F???

"width:500px;margin-left:-250px;left:50%;" what the #$#$@?

Are they kidding?

Wow, I didn't actually know that. That feels...so dirty!

Also utf-8 (or any Unicode) wasn't supported by most (all?) Netscape versions. Falling back to default Latin1 works with English text though.

Try browservice: https://github.com/ttalvitie/browservice

Reply on news.ycombinator.com