for various reasons mainly relating to my work i think i should read more often, and the only way that's happening is if i post about it.
2023-09-22: read a few books over the last little bit here.
Mystery of the Yellow Room. a classic of mystery for a reason. excellent.
the decagon house murders. a fascinating mystery in the vein of and then there were none. It feels a bit unfair in places but it's excellent in the end.
2023-08-21: Read more of the Network+ test study guide over the last couple of days. Yesterday I got through the chapter on VPNs and secure network protocols, and today I got through the chapter on IPv6. Setting up IPv6 networks seems like more of a pain in the ass than setting up IPv4 and the notation standards for it seem a lot harder to read than IPv4. Why can't the IPv6 notation just have a : for each part you're skipping over. Why is it allowed to use one :: to say an indeterminate number of zeroes has been skipped, but only once? Why is an address shortened to something like 2001:5c0:1000:a::29f and not 2001:5c0:1000:a::::29f so that there's less room for ambiguity when trying to read the address? Like, I understand some parts of it but it feels like if you're going to shorten the address to be easier to read you shouldn't have the equivalent of ellipses in it. IPv6 seems just a bit less clear than IPv4 does. I'm about six chapters from the end now, and then I'll probably go and work on brushing up on the concepts with practice tests to try and drill the concepts down a bit more, and also set up a few basic networks to drill what it actually means into my head, and then try taking the test proper.
2023-08-19: I just now remembered I have this page. Part of the reason I've been less active on the rest of the site over the last couple of weeks is that I've been spending most of my non-posting free time reading a study guide for the CompTIA Network+ exam instead of playing video games, since I'd like to switch careers and knowing how to fix networks seems like it probably won't get automated away. After all, if some AI can't access its training data because a cable got unplugged somewhere then it can't exactly scrape the internet to tell you how to fix the problem.
A lot of it's just been the sort of historical/fundamentals so far (talking about the OSI Model and the TCP/IP model and talking about the pinouts and speeds of the various CAT# specifications and various logistical aspects of the work, like making sure you ask for a floor plan before running any cable, making sure you are allowed to run cable in the walls, that sort of thing). After that it's been a few chapters going over some of the major network ports. I think it said the port areas are divided into 3 main zones between 0 and 65535. The common area was something like 0-1023, the middle area was some weird range like 1024 to like 3*2^14-1, and the ports that the end user can play around with were like 3*2^14 to 65535. dunno why they didn't just go with powers of two all the way down but I am not IANA. Actually, checking the ports, it looks like the range there was defined as 2^14 + 2^15 − 1, which would work out the same as what I remembered it being, but it's a bit more rigorous than what I was getting when I was trying to figure out how it divided by just dividing it by two until I ended up at an even 3. Also I guess the proper terms are "well-known ports", "registered ports", and "dynamic ports". The main ports it's covered so far are FTP (20, 21 iirc), SSH (22), Telnet (23), SFTP (115), TFTP(69), a few of the email ports like POP3 and IMAP (110, 143) and a few of the name resolution ports that I'll need to reread a couple of times because all the BGP/UDP parts bounced off, aside from the part where I just remember it came about as a way of extending the lifespan of IPv4, with added network security as a bonus. I can't remember all the details of how it works, I think it translates local IP addresses on a network into a unified IP address on the rest of the web, keeping some alias in the packet so that the border gateway can direct it back to the right address on the local network? Something like that.
And that reminds me of the chapters going over how the subnet mask works. My eyes were glazing over for a lot of it and I'll need to reread it. I think you set some local IP address like 192.168.0.1 or something and then set the subnet mask to some number that works out to a line of consecutive 1s, so like you could do 255.255.255.0 or 255.255.0.0 or 255.224.0.0 or 255.255.248.0 or something, and then you have however many zeroes of space minus two or something that you can add as a part of the network.
Today the chapter I'm reading is talking about DNS and the hosts file and namespaces. It all seems very boring, but from what I've read from various techs online, DNS being configured wrong is one of the most common problems. And, reading further, it does seem to address that. Or at least, it says there's usually some problem with the TCP/IP configuration, and then it gives some troubleshooting advice that goes like this: Check the NIC, check the NIC's driver, try pinging other local systems, check that the IP address and subnet mask were entered properly (things break here a lot), try to renew the DHCP lease (call the server admin if this fails), run netstat -s and see if some application is failing, try pinging the router's IP address (both local and the one connected to the internet), ping to a website (use tracert to see where the packet stops).
2023-07-08: finished reading maltese falcon by dashiell hammett. it's solid enough but I like it less than red harvest. sam spade's what the worst person you'd ever meet fantasizes about being.
2023-06-16: read an act of comedy of errors to see if my opinion on the bard's changed much in the last few years since I last tried making sense of what people see in his stuff. I think I've softened a little bit on it, but I still don't much like reading it. maybe next time.
-read the dain curse by dashiell hammett. it was published in parts in a serialized magazine and it feels it. the first two parts seem different than the last two, being a more generic detective story with a phenomenally stupid parlor scene (no mr. detective he didn't kill her and i didn't kill her, i had the toddler kill her!) and a supernatural mystery that has more parallels with professor layton than a hardboiled crime novel respectively, while the back half returns back to earth and hammett seems to have found his footing again, probably after a negative reaction to the experiments in the first couple of parts. it's a bit weaker than red harvest all-around, but it's still a fun read. I'm really interested in seeing how hammett's style continues to develop in the rest of the books.
-read another chapter or so of the three musketeers. d'artagnan has gotten himself embroiled in a duel with each of the three musketeers on the same day.
2023-06-14: read red harvest by dashiell hammett. it's a really fun read but boy there's a lot of stuff in there that gets concerning when I think about how the author worked for the pinkertons and the detective is just a thinly-veiled pinkerton agent. all the little justifications he makes to himself about how the best way to solve the case involves a lot of murder and how he's responsible for more killings than any of the people he's investigating sort of flies in the face of any moral high ground he might have had any claim to. absolutely wild pinkerton apologia here. still, it's a really fun read.
2023-06-01: listened to three chapters of the three musketeers. It's still in the middle of setup and I haven't seen all four of the three musketeers yet. it seems much more brisk than the count of monte cristo.
2023-05-26: read the first three or so short stories in the early cases of akechi kogoro by edogawa ranpo. the first two were pretty weak but the third one started feeling like it had more of a gothic horror mood to it and I think ranpo is starting to find his footing. interested to see how the voice develops as I go through his catalog.