- 3 Posts
- 57 Comments
1·7 days agodoesn’t he weasel out of the responsibility to give clear, logical, verifyable reasons for his position?
Absolutely, if I remember right he leans back on having experienced bad comments more often than helpful ones. John questions this. I think it is close to dogma with Bob on this.
Can you explain that more?
And doesn’t the example with the prime number generation algorithm show clearly that omitting context just does not work for code?
Quote from https://en.wikipedia.org/wiki/High-context_and_low-context_cultures
High-context cultures often exhibit less-direct verbal and nonverbal communication, utilizing small communication gestures and reading more meaning into these less-direct messages. Low-context cultures do the opposite; direct verbal communication is needed to properly understand a message being communicated…
Now I’m not making a strong claim that Bob and John are from different ends of the context spectrum. However it seems to me that Bob believes there is enough ‘context’ available in code and in coders themselves to communicate all meaning without comments.
Even Bob’s diagram, to help explain the primes algorithm, assumes high context in the reader. It’s lacking any labels or key - we are just supposed to see what he means if we stare hard enough at it. If we are already immersed in the problem space then this might work but its so inefficient for anyone else.
And once we step away from our code for even a short time we are that someone else. We are going to waste a lot of time rediscovering how the algorithm works. A case John makes convincingly I think.
Code cannot replace comments. The primes algorithm avoids division I believe but this is not clear from the code alone. A reader might work this out eventually but a comment saves so much time. Could the code be refactored to clearly express the avoidance of division? Yes there’s probably a way, but imagine how bad that code would read and what a waste of time just to avoid a comment.
Stow/chezmoi/your choice for dotfiles, config mgmt for system config. You don’t need to rebuild whole server to start with ansible tho, you can take over one file at a time and grow as you learn.
As you’ve found I don’t know of a tool that will cover both usecases as config mgmt for dotfiles is too much and dotfile mgrs for system config is probably out of their scope.
1·11 days agohttps://blog.fyralabs.com/upgrade-error-with-vlc-plugins-freeworld/
That rpmfusion package looks like trouble, can you uninstall it or all of rpmfusion before upgrading? I also see internet problems here also so you might be better off fixing that first.
Another thank you for posting this, made my day.
I have read and followed a fair amount of Uncle Bob’s work but was not aware of Ousterhout till now. Bob says during the time the Clean Code book was written there was an anti-comment sentiment about and this matches my own experience. I agree with Ousterhout that it’s taken too far in his book though.
I wonder if there is another factor at play - some people/cultures prefer high context communication and some less. Bob seems clearly to prefer low context i.e. the burden is on the (code) reader to educate themselves. Whereas John makes it a matter of professional behaviour that we make the next reader’s work as simple as possible by commenting code as appropriate.
Surely it’s better to assume high context is needed and provide it (within reason) versus only catering for low context. As Bob discovered he became a low context person when he returned to his own code after some time had passed.
8·11 days agoMessages are only sent when both online though, thet’s the bigger difference (unless using Briar Mailbox). Also it can send over wifi and bluetooth without internet connection i.e. no other devices involved.
Upgrade from compression tools to backup tools. Look into using restic (a tool with dedup, compression and checksumming) on a filesystem which also checksums and compresses (btrfs/zfs) - that’s probably most reasonable protection and space saving available. Between restic’s checks and the filesystem you will know when a bit flips and that’s when you replace the hardware (restoring from one of your other backups).
If you mean with a gui then this is one tool:
Try this:
gsettings set org.gnome.mutter check-alive-timeout 10000
Value is in ms so 10000=10s. You can also set it to 0 to disable.
To be clear you can absolutely type in exact measurements in the ‘tool controls bar’ for all the shape tools and even individual nodes. Scale and units are easy to switch between and there’s a measurement tool also.
I know. The author suggests:
Experiment with new-to-you version control systems like Fossil, Mercurial, and Pijul.
The author is:
learning about different version control systems. For example, the differences between Fossil and git revealed a lot of my biases towards git simply because it’s familiar (and Fossil seems really cool). Reading about the theory behind Pijul absolutely bends my brain into knots. I keep trying anyway because conflicts in git are frustrating and I’d like a better solution.
The author says:
It would be nice to move beyond git one day and have a better experience for managing complex codebases, and not on GitHub’s timeline.
Jujutsu is another git alternative I keep seeing around and came to mind reading this:
https://steveklabnik.github.io/jujutsu-tutorial/introduction/what-is-jj-and-why-should-i-care.html
From the zdnet article linked in another comment:
tech is one thing; business is another. That’s where the RSL Collective comes in. Modeled on music’s ASCAP and BMI, the nonprofit is essentially a rights-management clearinghouse for publishers and creators. Join for free, pool your rights, and let the Collective negotiate with AI companies to ensure you’re compensated.
I guess this is the body that will be leading the enforcement/bringing the consequences
Sorry re-reading my comments it’s not super clear what I meant: nowhere else in the table do they take account for the ‘hidden’ on-going maintenance of looking after a server/self-hosting. So this is the only row where they address ‘cost’ and I just thought it’s a bit optimistic to say replacing all of Spotify just costs a one time server setup and storage. I think you’re saying this row was only meant to indicate financial cost and I agree it’s basically accurate from that meaning. However this is only the ‘initial’ cost. For example a self-hosted server and storage will eventually have to be replaced whereas Spotify will just keep replacing their own servers and that’s already baked into the price of your subscription (caveat: that Spotify price will rise over time).
It’s not a big point really, maybe I’m nitpicking.
I see what you’re saying but nowhere else in that table is cost mentioned. Below the table they say maintanance is minimal. If you’re already looking after storage, containers and server(s) I guess that could be true.
Author says “one-time server setup + storage” but there are a few moving parts and always updates to handle so I’m sceptical this could be truly called ‘one time’ (or any selfhosting). Time will tell I guess. I enjoyed the article though and gave me food for thought.
10Gb is not a big file relatively speaking - both ext4 and btrfs (for example) can handle 16TiB and larger. If this was your only reason for choosing exFAT then you can definately migrate.
Are you using RAID at all? If so ZFS is probably the way to go. If not I think it matters less whether you use either btrfs or ZFS.
Regarding btrfs and power loss:
…btrfs is designed to only experience data loss not corruption, assuming well behaving hardware in power outage scenario. In practice ZFS has more maturity overall (definately) so may be better (my speculation).
Beyond direct comparisons if you already have on and offline backups then you are protected from power corruption and only have to worry about data loss anyway?
Was this comment meant for a different conversation? We’re talking about VPNs here.
Yes it’s not too much bother to set this up, it can be put into ansible and once working I’ve not had to touch it again. Here’s another dracut tool using dropbear that works well and has decent instructions on setup: dracut-crypt-ssh