For those looking for PDF, it is at the author's site (http://www0.cs.ucl.ac.uk/staff/M.Handley/): http://www0.cs.ucl.ac.uk/staff/M.Handley/papers/only-just-wo...

As an aside, our server may well be the oldest still running on the Internet:

  telnet www0.cs.ucl.ac.uk 80
  Trying 128.16.6.8...
  Connected to haig.cs.ucl.ac.uk.
  Escape character is '^]'.
  HEAD /staff/M.Handley/ HTTP/1,0

  HTTP/1.0 200 Document follows  
  MIME-Version: 1.0
  Server: CERN/3.0
  ...

Proudly running CERN 3.0 on a Sparc 20 continuously since ~1993.

Continuously as in No Downtime or as in you've just been using the same machine that long?

That's... gonna depend on how popular a certain HN post becomes...

Good answer. The Slashdot Effect delivered by its best alternative. ;)

Welp, here we go.

It has definitely been down occasionally - the odd power outage guarantees that in 23 years, and it doesn't really merit a UPS. Also I think it's on its second Sparc box, but our sysadmin has a pile of old spares. And it had to be patched to be Y2K-compliant, as the CERN 3.0 httpd was last updated in something like 1994. It is definitely slow compared to modern systems, so it's no longer our main webserver, but it still handles the occasional HN load spike reasonably well.

That's along lines I was thinking. Haha.

Oh man, I emailed the creators a year ago when they first released the site, encouraging them to make the frontend open source so the performance and layout issues could be fixed by the community (including myself, there are some easy bug fixes I wanted to submit). Back then, they were still deciding what the direction was going to be for the library. A year later, it seems they've opted for the closed-source, no improvements path unfortunately.

I still subscribe, since the papers are interesting, but the website needs some love.

This seems a bit harsh. As far as I understand the purpose of this website is to make papers easier to understand. I think it accomplishes that. They are not even trying to show you ads. how are you helping?...

Also, they might have restrictions about distribution of papers. Youtube also doesn't have a download link next to every video.

With all due respect, perhaps it might have been interpreted as harsh, but what was noted above is no less true.

I'll acknowledge there is not an overload of ads being displayed with the paper, which is cool. Also acknowledge that there certainly might be restrictions for distributing papers (a separate, possibly inflammatory topic)...but have you tried doing something as simple as printing the paper? On Chrome (windows machine) it prints out so disjointed as to diminish one's interest in reading the paper (offline of course). I just want to print this out and read it on the train ride home. I guess the intent of the site is ideal, but execution is less so.

It printed mostly fine for me (mac, chrome). It did not print the comments, which to be honest I'm not even sure about how you would go about showing the comments (which are often very large) in a printed page.

"Thanks, for me this was mostly another illustration how to make a website that doesn't." - sounds pretty harsh to me.

Well, but the site just doesn't. Not harsh, just truth.

I was unable to read the paper because an unclickable modal kept popping up asking me to subscribe.

Ignoring any arguments about tone, I also agree that not providing at least a link to the actual paper is a puzzling omission. If the goal indeed is to aid paper comprehension, limiting the ability to read it is a weird choice.

And not the gp, but I'm not "helping". In fact, I have no idea what I'm supposed to take away from you having asked that question.

It might sound harsh to you, but you'd feel differently if you sat in front of my computer. Each misclick costs me 1-2 seconds of layout recalculation (100% on 1 core). For reference, my computer is AMD X250 from ~2011, 4GB RAM (enough), and running Firefox on Debian 8.

Using chromium it's a tiny little quicker, but still annoying. They could easily stop the content bumping and the darkening "please subscribe" overlays.

Try disbling Javascript, I disabled it by default and the site is really fast for me and I can read the document just fine. I enabled JS and have to agree that now it is pretty sluggish.

I don't understand the need for expensive animations or the imagined use case of opening/closing the comments sidebar a hundred times while reading the paper. Why does opening one note hide all indicators of other notes? (It seems if you're reading at above ~1500px browser width, it shows them as small dots, but removes the preview text of the note). I should just be able to leave it open and have new notes/comments appear as I scroll past the relevant section.

Why does clicking anywhere on the document, or highlighting any text open up a "write a new comment" sidebar? Do they really think there are going to be more users writing comments on a paper than there will be copy-pasting a few sentences from it?

It feels like the primary purpose of the site is for users who want to comment on papers, and not users who just want to read papers. For the latter group, it's a really poor experience.

well I have a mbp late 2013 and it costs me 1-2 seconds (16 gb ram and core i7 3rd gen...)

sure. talking to people on the internet only just works...

This was probably a witty reference to the article. Can't appreciate it since I can't bring myself to suffer the website...

I loaded it up on my mobile page to see what I thought was a PDF reader opening up to find it was a browser with PDF covering the whole page grayed out with a little box at the bottom-right with a little X or something. I can't remember as I closed the page immediately. It was maybe the 2nd time I've seen something like that used. I knew someone would post an actual link to the PDF or a site less broken in presentation.

So, yeah, it was an unusually, negative experience on mobile.

The implementation is plain broken. I can't scroll at all with my trackpad. Every time I click anywhere, a "Comments" window stutters out from the left side, and takes a few seconds to retreat next time I click.

It also have comments pane which takes half of the screen (with non working X button) and hijacks back button (firefox on android).

It is really that hard to just serve PDF instead of stupid javascript filled browsers.

> It is really that hard to just serve PDF instead of stupid javascript filled browsers.

Well, the entire point of the site is a collaboratively annotated journal article. You may not think that's interesting, or that it's feasibly implementable with current tools, but I think the folks do a pretty good job with their stated goal. So I think this is something to be left up to the HN voting.

It worked fine with NoScript -- which is a good way in general to hide from the madness of the modern Web.

Agreed this is just awful