Making a Popularity Contest out of my Website
A few of you may have noticed that I've just added a page detailing the 'popularity' of all my articles - in other words, how many 'GET' requests (individual downloads, from server to client) they have had.
It was something I had thought about doing for a while, and decided on a whim to do it today. It was easier to implement than I expected, taking an afternoon, and runs surprisingly quickly.
Being based wholly on user-agent information, there is no real privacy concerns to be seen here (as it is simple enough to spoof your user-agent and use Tor if you are concerned), it only tracks the times a page has individually been downloaded. The script may come to be useful for people looking to achieve a similar thing, although it is long enough it would be impractical to just paste and describe all of the script here.
The script looks for articles (I do not have to tell it which links to look for manually), uses grep to see how often a GET request occurs for each file in the nginx host logs, and does very basic arithmetic to calculate the size of bars to eventually display as HTML.
Most of what I found made sense, and has reassured me that everything is working correctly.
For example, my post on making a Tor mirror site as expected had more Tor hits than clearnet, although Tor users make up a minority on my site everywhere else. My latest article on 'Tetsuo: The Iron Man' is the lowest ranking, because it is new and has not had much traction yet.
This script will update nightly, and it should start ranking this post also. It should be interesting. That's all for now.
All my Sites are Available on Tor
It's been a while. The longest I've left it actually, about a month or so. I make no promises to improve this track record, because I prefer quality to quantity, but it is good to be back. Life has been quite chaotic for some time.
Is there life on Tor?
My sites are now on the Tor network, and any further subdomains I make create will also be on the Tor network. For those less familiar, it may seem strange to host a site on the clearnet and associated with an IP address already as a hidden service, but there are anonymity advantages to never leaving at an "exit-node" and remaining on the Tor network.
And now, the links:
The sites are largely the same as their clearnet mirrors. There are two differences you may note with concealed.world - I don't bother to report on IP-based analytics (I can't see your IP), and I curate the Lainchan webring links to bring more focus to sites also on the Tor network.
As for git.concealed.world, yes, you can clone repositories directly from it. You would probably need to use torsocks, or something similar.
I won't make a separate post about how I set this up, as it is stuff you can all find fairly easily on the internet. I will point you to Tor's own documentation on running services and this "best practices" guide.
All the other work I have done is creating shell scripts for far easier housekeeping - rsync, chown, chmod, and dealing with systemd - but this is all the kind of thing you've seen elsewhere from me. As always, it's much easier to run something like 'sudo onion -c' than deal with a whole bunch of rsyncs and permission issues manually.
The dates on articles are wrong
You may also note that the dates on my articles no longer reflect when they were posted. If you were online around 9 PM GMT on the 8th (unlikely, but I'm sure you found it hilarious if you were), you will have seen that I accidentally deleted my whole site.
Yes, don't ask. I'd been doing a lot of fiddling at the back-end. I'm ashamed of me too.
Thankfully I had kept backups of most things, and only had to retype a few articles from my newsboat cache, and check it over in paranoia for a while.
I had never tried to claim that the dates were anything more that when an article file was last modified, but it's still annoying. For archival purposes, I'll mention this site got its first post on the 6th of July 2020, was first fiddled with over April (before I got distracted by video games), and posts following that were made in groups every few weeks. Based on the when I pushed to my Git server, you can tell approximately when I was active.
Automating Writing HTML
What kind of psychopath goes as far as to automate writing plain HTML? Well, me apparently.
At first I didn't think that going this far for automation would be worth it. It was actually my friend over at Saltorn that had suggested the idea to me, seeing as he can automate such things with Emacs when writing things of his own. I thought it would be too far to go in the name of automation, and that the returns would be diminishing.
But in actuality taking out more of the manual effort in writing allows a lot more free flow of consciousness, I tend to find I write more quickly and easily now.
You can find the script here. Although it's suited to my particular usage, you could quite easily adapt it yourself if you're looking to do something similar. I realised that the components I use in my articles could easily be boiled down to a dozen or so different elements, which I could write quick shorthand for.
The elements start consistently with a '@', which is a character I would rarely use, and end consistently with a ';'. Using non-frequent characters means it's less likely there will ever be a case in which it breaks.
And yes, I did write this article using the script, too. It's nice.
Fighting with Web Standards and the Scrollbar Fiasco
It has been a while. About a month, precisely, although you can see that I've still been alive due to the new posts on my music list. I hope you didn't think I had abandoned you yet.
And as I have done before and will certainly do again, I've got a backlog of posts to spam, that will probably come out over the next couple of days, on various subjects. (For example, the fact that I am writing this post in my own crude markup language, rather than plain HTML).
But first and foremost, to the most immediate change, and in reference to the title of this article. I've been having a back-and-forth with Qorg in regards to this - I like my website without a scrollbar.
A while back it was brought to my attention that this doesn't work with Pale Moon browser - it isn't entirely up-to-date with its CSS seemingly, and if I hide the scrollbar, it is impossible to scroll. I initially thought I had resolved this, although the fix was rather crude, relying purely on the user-agent.
Eventually I needed to create a solution that would work for those that use Pale Moon, but understandably (for privacy, and to avoid dealing those such as Cloudflare) spoof their user-agent. They would not be able to browse my site. And any browsers I did not know of that had the same issue, would also not be able to browse my site. And what if you just preferred to browse with the scrollbar?
So, I have now created the option to browse with a scrollbar, dependant on a single first-party cookie. See the change here. You can visit this page begin using this site with a scrollbar. I personally don't prefer it, but it's not a bad thing for people to have the choice.
My Personal Music List
So, I listen to a lot of music generally, I have opinions on it, I keep a pretty sizable collection. And I thought to myself, "Hey, why don't I make a post for it? It would make for some good content - if I didn't all that knowledge just sits there."
How wrong I was.
It has easily become the largest plain HTML file on this entire site by far. Vim screams in agony. I didn't quite realise I had a problem until I decided to type it all down - and then realise how long that could take me.
Some Statistics (at the time of writing this)
- Number of rated albums: 156
- Total items: 360
You get the picture. Once I had started it was far too late to stop, my fate was sealed. I do hope however, it is of some value to you. Go check it out:
A Therapeutic Mini-Rant About Cloudflare
Normally, I try to avoid reiterating things others have already done. It's not worthwhile to me to get attention by echoing things other people have already said, especially if I haven't learned about it any further through my own experience.
But the reason I come here with this today is that I have even noticed fairly small web projects from independent developers using Cloudflare. It's not even big corporations any more. And I'm not strictly sure people know much about the potential pitfalls, in regards to internet privacy and centralisation. I certainly hadn't read up much on it until recently - all I knew was that Cloudflare left a bad taste in my mouth.
To sum up the basics:
- Cloudflare discourages the use of anonymising networks, by making websites difficult to access.
- Cloudflare also makes people on poor internet connections have great difficulty in accessing websites.
- Cloudflare is a centralised gateway to the internet - thus by using it, it is exceptionally easy to profile you through its logs on you, and when it has problems, everything goes down with it.
Sources on the privacy of Cloudflare, and how it works on the back-end, aren't terribly accessible. I encourage you to do your own research. But for a starting point that somewhat reflects how I feel about it, take this.
Cloudflare is a piece of software that operates firewall rules, classically to protect your host from threats as such DDoS attacks. There are other options to do this - and even better, you can simply learn how to use iptables yourself.
In short: if your website requires me to connect to Cloudflare and complete a captcha, I'm not going to visit it. If you're already using a website with Cloudflare, seriously consider your use case and whether it is absolutely necessary for you. I would hate to see people acting to centralise the internet without giving it much of a second thought.