David Zaslavsky

Software engineer, former particle physicist, occasional blogger. I support the principle of cake.

Public Key

npub105vywj5u9qx5fkvuua6cq6xhmm2qmnltetnl0pqzk3wjn452fphsw0tsfs

NIP-05 Address

diazona_at_techhub.social@mostr.pub

Profile Code

nprofile1qqs86xz8f2wzsr2ymxwwwavqdrtaa4qdel4u4elhssptghff669ysmcpz4mhxue69uhhyetvv9ujumt0wd68ytnsw43qykn97q

Publishing to

relay.mostr.pub

Author Public Key

npub105vywj5u9qx5fkvuua6cq6xhmm2qmnltetnl0pqzk3wjn452fphsw0tsfs

Show more details

Last Notes

2024-11-06 00:48:02

- reply

@npub1chs…ezjw whoa I have a friend from high school who lived there for a while, small world
#distractodon

2024-11-04 03:00:16

mandible hug
#Monsterdon

2024-10-20 20:59:13

- reply

@npub14vg…asrr Only 50 bits though, whereas RSA keys used for SSH are often 4096 bits - I don't know about other protocols where RSA is used, but I'd guess they're roughly comparable? So it's a significant step forward, for sure, but going by the article, it sounds like we've got some time before they scale up to breaking "real-world" encryption.

2024-10-05 22:10:03

- reply

@npub1na3…jx6x oooo this is cool!

2024-08-19 04:37:30

- reply

P.S. I hope y'all appreciate what I had to put myself through to create this, skimming through the ENTIRE MOVIE *AGAIN* to find suitable screenshots and audio clips 🤦 🤯 🤦
#Monsterdon

2024-08-19 04:35:40

Tonight's #Monsterdon in a nutshell 🐙
https://files.techhub.social/media_attachments/files/112/986/820/592/365/780/original/d4ccf67dae692a99.mp4

2024-08-19 01:58:21

- reply

@npub1jpp…u59l If last week's movie is any indication...
#Monsterdon

2024-07-08 22:42:30

- reply

@npub1pfe…e2dm In the not-so-distant past I've run into a site at work that has exactly that problem on its login page, such that our own gTLD fails validation 😂

2024-06-18 03:03:07

- reply

@npub178r…qekt I kind of wanted reference 4 to be the Lock Picking Lawyer's channel page 😂

2024-05-02 22:50:02

- reply

@npub1pfe…e2dm "simple" and "intermittent failure" in the same sentence, that's ambitious 😂

2024-04-24 21:24:38

- reply

@npub1km9…rh4n I usually come here to get away from outrage so yeah, easy decision

2024-04-24 21:20:47

- reply

@npub1km9…rh4n ooh I see.
I'm gonna guess this is not worth my attention though 😇

2024-04-15 10:47:37

- reply

@npub14zy…2z57 interesting, although not at all what I expected when you said "compatible with Python" 😛
Interestingly, the Erg documentation is only the second place I've ever seen "NG" used to mean "incorrect"

2024-04-04 05:47:16

- reply

@npub1km9…rh4n I'd want to do the same, if I can get time... If you happen to write up a blog post or anything about how you do it, I will be quite interested 🙂

2024-04-04 05:25:05

- reply

@npub1km9…rh4n Yeah, I think I've seen a couple people do that. Sounds like an excellent idea to me!

2024-04-02 00:40:12

- reply

@npub1pfe…e2dm haha definitely 👍

2024-04-02 00:34:28

- reply

@npub1pfe…e2dm what am I looking at 😂

2024-03-13 22:29:00

- reply

@npub1gaj…78dz Haha yeah sounds like a good plan! I definitely cannot account for how a screen reader complicates these things. (I would expect it's possible to change the shortcuts in nano but I never looked up how to do it.)

2024-03-12 23:25:05

- reply

@npub1mjj…h2pr Huh, weird. This definitely sounds unusual.
Anyway, the main point I wanted to make (which I kind of forgot to get to before) is that it's an overreach to say that the process is "beyond horrible" based on your single atypical experience. It still works fine and smoothly for a lot of other people.
I suppose you could share the errors and maybe get some help here, if you wanted to.

2024-03-12 23:15:34

- reply

@npub1mjj…h2pr For plain old Git, at least, I never found it particularly hard to handle. I guess figuring out how to generate a new SSH key if you don't already have one can be a little tricky? But it is just one command to run (at least on Linux, or I think no more than a couple button clicks or such with typical SSH clients on Windows), and beyond that, all you have to do is take the content of a file, paste it into a web form, and you're all set.
Not that this does much for you, I suppose, except to suggest that if you're having this much trouble with it, there may be an easier way because I don't think it's supposed to be that hard 🤷

2024-03-03 04:45:56

- reply

@npub1jeq…jglm Well yeah I knew it had to be something other than Mastodon (or at least something other than stock Mastodon). I didn't realize there was any ActivityPub-compatible software that supported custom formatting though. The only one I'd heard had started experimenting with it was CalcKey, and that seemed like a pretty big deal so I thought there would have been more buzz if another one came on the scene.

2024-03-03 04:33:22

- reply

@npub1jeq…jglm Ah yeah, I was using the Android client to view it earlier, but now that I look on the website it seems fine. Maybe the Android client can't handle numbered lists.
Incidentally, how did you get a numbered list in the post? Some kind of custom server software I imagine? (This is just out of curiosity)

2024-03-03 04:19:40

- reply

@npub1jeq…jglm BTW I dunno how it is for you but I see a bunch of missing line breaks in your post

2024-03-02 05:40:56

- reply

@npub1gaj…78dz Oh interesting, I didn't think emailing was needed either - I ran mail servers on my Linodes for many years and they always seemed to just work.
I guess that restriction could be something newer that they didn't retroactively apply to old instances, that would explain it 🤷

2024-03-02 04:54:10

- reply

@npub1gaj…78dz Oh nice, congrats 🎉
Can't say I'm surprised, honestly... Seems like you never know what you're going to get on a marketplace like that. I would also roll my own if I were doing this.

2024-03-02 04:15:17

- reply

@npub1gaj…78dz Best of luck debugging this, but it seems like the kind of thing where it might be hard for someone else to help without getting pretty hands-on - like, at the level of being able to connect to your specific server, or something akin to doing a screen share and watching you run commands and such. (I would offer suggestions if I could think of anything)

2024-02-29 18:30:56

- reply

@npub1gaj…78dz No worries, sorry for confusing you! I had intended the laughing face emoji and "but seriously" to indicate that it (the part about Toastmasters not working) was just a joke.

2024-02-29 17:35:23

- reply

@npub1gaj…78dz oh I was making a joke

2024-02-29 06:06:22

- reply

@npub1gaj…78dz If "wow" is all you have to say then it didn't work very well, did it 😂
but seriously, good luck and enjoy! Sounds like fun.

2024-02-23 01:00:55

- reply

@npub17qe…v7la Hmm interesting, I'll be keeping an eye on this. If I had thought Bluesky would ever do this I probably would have been more open to joining it a while ago.
I gotta say, whoever wrote that blog post seems confused about how certain things work on Mastodon. (Like this idea that your server determines your community)

2024-02-06 01:57:01

- reply

@npub1e7v…flg9 Unfortunately I am more confused by this 😕
Like, for one thing, when you say "there is no package, I'm installing the top level application" - how are you installing an application if not by using a package of that application? (or something that could be built into a package, like a Git repo)
(No obligation to explain it, of course; I'm not here to consume any more time or effort than you want to offer, and if that means I just won't understand, that's fine.)

2024-02-06 01:47:50

- reply

@npub1e7v…flg9 Hmmm... I'm having a hard time understanding the situation where this fails (assuming `metadata` is `importlib.metadata`). I don't suppose you have an example to show?

2024-02-02 18:44:31

- reply

@npub12ze…hunm Vivaldi is one of the many derivatives of Chromium, so if you care at all about Google not having a monopoly over rendering engines, that's a big point in favor of Firefox. 🤷
Other than that, I guess it comes down to personal preference? I mean, I prefer the UI and experience of Firefox, but you can really make your own decision based on that.
#Firefox #Vivaldi

2024-01-30 20:54:48

- reply

@npub1qag…9gyn No you're not. What you're saying is at odds with every relevant understanding of the word "link" I'm familiar with. That being said, as far as I'm concerned, it doesn't really matter whether your definition of "link" is (accurate) fact or opinion or wrong. I understand what you meant in your original toot, which is what matters.

2024-01-30 20:42:13

- reply

@npub1qag…9gyn Ahh yes I see where you're coming from now. I would disagree that "the link is not the link plus tracking information". The link is whatever appears in the href attribute of the HTML tag. That's why I was confused; your definition of "the link" as something different from the unmodified contents of the href attribute was throwing me off.
I fully agree that it'd be great if copying the URL without tracking information were the default/easiest option and users had to jump through extra hoops to copy the unmodified URL that includes tracking info. It's just that I'm a little uncomfortable with that action being labeled in a way that suggests it's copying the unmodified URL, such as the wording "Copy Link". I dunno, maybe it's worth the tradeoff, but I'll never be convinced that it's not at least somewhat inaccurate. (Plus I could see myself being very confused by that behavior in practice.)

2024-01-30 20:08:17

- reply

@npub1qag…9gyn That makes sense. And that is what I would have expected you meant based on the normal meaning of "opt-out", but on my first reading it seemed like different parts of your post might be inconsistent about what they meant by the term, although I assumed it was because I wa misunderstanding the meaning of some of those parts, I just wasn't entire sure which ones.
After rereading I think the part that confused me is this: 'The simple "Copy Link" option should do what it says, it should just copy the link and nothing more (opt-out=default)'. Specifically, as I understand it that's describing exactly what Firefox does, but something about the wording makes it sound like you're asking for a change. I'm also not sure what the "(opt-out=default)" at the end is doing there.

2024-01-30 19:37:04

- reply

@npub1qag…9gyn I'm a little confused here, what do you mean by opt-out in this case?

2024-01-29 03:26:08

- reply

@npub1jpp…u59l wat that's actually a website? 😂

2024-01-25 23:08:47

I just spent 3 days debugging why a CSS rule wasn't working only to realize it was because I hadn't written the rule in the first place 🤦😂
(actually #Python #Textual CSS, not normal (browser) CSS, but the point stands)

2024-01-18 17:38:08

- reply

@npub1e7v…flg9 This sounds like the kind of situation for which we (@npub14tk…0vpx and I) created https://github.com/diazona/setuptools-pyproject-migration 😉
I mean, if you're just complaining, that's cool, but if someone takes this as practical advice, I would want them to know there is a better way
#Python

2024-01-04 01:41:49

- reply

@npub172v…689n Nah I didn't mean to request a blog post (I'm not the type to make such a request so obliquely), but if you did feel like writing one it might help clear up some things. Like, when you say not to use namedtuples in new APIs, is that your opinion or are you representing some kind of official position (which is what I thought you meant)? Or, if it's not official, is it an opinion widely shared among core devs, such that usage of namedtuples might be deemphasized if not actually deprecated in the future? In which circumstances are namedtuples appropriate? And ultimately the important thing is probably, why are namedtuples otherwise undesirable and what would you suggest people use in cases where they're not appropriate? Like, would this new proposal you have be a suitable namedtuple replacement, or would you suggest dataclasses in some cases, or something else?
No obligation of course, I'm just trying to represent my perspective of what information I'd find useful.

2024-01-04 00:31:10

- reply

@npub172v…689n I feel like that message needs to be much more widely publicized. I've seen one or two blog posts or the like arguing that dataclasses are better than namedtuples, but nothing even semi-official discouraging their use.
#Python

2023-12-27 20:11:02

- reply

@npub172v…689n Personally I wouldn't see myself using this new type. I mean, we already have named tuples and dataclasses which serve a similar role, and I already find the existence of those two types to be a source of uncomfortable ambiguity - not that they're the same, of course, but I often encounter situations where I could use either one and I have no good reason to choose one over the other, and the existence of record types as a third option usable in many of the same cases would only increase that ambiguity.
I suppose if other people find it useful, who am I to object - it's not really hurting me to have another type syntax that I never use - but I guess I just don't get it 🤷
#Python

2023-12-25 04:32:10

- reply

@npub1c9a…knew I've spent many Christmases on my own (not really by choice, but it is what it is) and I've got a fairly well established set of tradition at this point:
- cook lasagna and brownies for dinner (except this year I ran out of butter) with my favorite sparkling grape juice to drink
- snack on some chocolates throughout the day
- watch a movie I like (often #TheMuppetChristmasCarol, best Christmas movie ever!)
- send some holiday greetings to my family
- otherwise just do whatever I feel like
This year I expect that to involve reading #WheelOfTime (I'm on book 12, hoping to finish the series by the end of the year) and listening to some good #ChristmasMusic
#JoinIn #Christmas #SitTogether

2023-12-25 04:24:29

Just realized I am nearly out of butter at home, which means my traditional Christmas brownies are probably not happening... it'll be fine, just a bit of a bummer after how much I was looking forward to them.
On the other hand it's one less thing for me to do on what is already a busy day.
#JoinIn #Christmas

2023-12-07 10:44:44

- reply

@npub1tx3…kxnn @npub1sg4…xslc Hmm... Well, each maintainer needs to do what works for them, of course, but personally the worst experiences I've ever had contributing to open source projects have been the ones that accept patches by email. I've been driven away from multiple projects by the sheer difficulty of understanding changes and following discussions in that workflow.
Some handle it better than others, but still, for me, no email-based workflow approaches the ease of use and clarity of something like #Github or #Gitlab.
#programming #Git

2023-11-27 02:25:24

Just when you thought smoking on a spaceship was bad, now they have guns and grenades
#Monsterdon

2023-11-20 03:09:59

I think you could have guessed what happens when you leave the guy yelling "you can't do this, this is madness" as the only one in the control room where the madness can be started and stopped
#Monsterdon

2023-11-09 22:08:38

- reply

@npub1pfe…e2dm Generally I look suspiciously at strongly-worded hot takes like this, but I do agree with the underlying point.

2023-10-21 04:11:37

- reply

@npub1gaj…78dz Well, I've confirmed that the button is there for me 🤷
https://files.techhub.social/media_attachments/files/111/271/046/924/435/781/original/e0e2919e564e0c12.png

2023-10-20 23:23:03

- reply

@npub1gaj…78dz Huh weird. When I get back on my desktop computer I'll check and see if I still see the follow button at the top of the hashtag search results.

2023-10-20 22:40:30

- reply

@npub1gaj…78dz If you're asking whether it's possible to do, yes it is (or at least was, I haven't done it in a while so I guess I wouldn't know if they removed the functionality)

2023-10-14 18:07:29

- reply

@npub1kpw…xzz4 Yeah that makes sense. This is why I hope Firefox catches up in that area. Personally I rarely have more than 15 or so tabs open at a time, so it's not that big a deal for me.

2023-10-14 18:03:56

- reply

@npub1kpw…xzz4 Yeah unfortunately #Firefox is lacking when it comes to tab grouping. I do hope to see some improvements in that area. But in pretty much every respect it's great for me. (My personal killer feature is multi-account containers)

2023-10-01 04:15:13

- reply

@npub1gg8…439g @npub1f66…xvfz @npub1kt3…5004 I involuntarily imagined that next-to-last one in Gonzo's voice 😂
Come to think of it... even though #TheMuppetChristmasCarol doesn't exactly have an opening line, the first few sentences of actual narration are a strong start:
(Gonzo) "Hello! Welcome to The Muppet Christmas Carol! I am here to tell the story."
(Rizzo) "And I am here for the food."

2023-09-23 23:39:27

- reply

@npub1k08…6xvn Is it claiming that it's all a scam or that these email systems aren't really private? Those are not the same thing.
TBH the whole premise sounds potentially kind of clickbaity to me, but I'm not interested enough to watch a video to find out. (If there were a written form of this, I'd definitely read it though.)

2023-09-22 21:34:01

- reply

@npub13vv…0sze Yeah it is a great feature, it's just that the downside I've found is that if you schedule a message for, say, the next day as part of an active discussion, then the message will be sent without taking into account all the developments that happened after writing the message but before sending it. So it arrives out of context, at least in a single-threaded discussion model like Slack uses. (it's less of an issue with branching discussion like in email, although that relies on people paying attention to which message is being replied to) That's my main reason for not using scheduling more often.

2023-09-22 20:33:51

- reply

@npub13vv…0sze Yeah that's true.
IMO in an ideal world, if you send a "please do this ASAP" message on Friday afternoon (assuming the typical M-F work week of course) you have to be prepared for "ASAP" to be Monday morning 😛 But of course the real world often doesn't align with that ideal... If it's really not possible to get people to accept that those emails will sit around until the next week, but you can get them to wait until Monday to send the messages in the first place, then yeah, that is a totally reasonable strategy.

2023-09-22 20:20:31

- reply

@npub13vv…0sze Hmm, interesting take. Personally I am of the school of thought that it's up to each person to manage their own inbox (and notifications etc), not up to the sender of an email to send it at a time that's going to be convenient for the receipt - though of course it depends on organizational culture and people have to reach some sort of happy medium.
I know about cultures where it's at least suggested to schedule messages such that they don't arrive too far outside of the recipient's working hours, but it sounds like what you're suggesting goes a fair bit further than that.
(there is not much of a point to this reply, I just find it an interesting topic) 🤔

2023-09-11 04:48:09

- reply

@npub15nw…s0l7 @npub1wct…h3tv Wait I'm kind of confused about what y'all mean by "export" in this context

2023-08-30 23:21:26

- reply

@npub1tx3…kxnn Ah, well... in that case, I appreciate that you accepted the contribution! Whoever wrote that pyproject.toml file seems to know their stuff.
I'd be happy to explain some things about modern packaging, and/or help you fix up pyproject.toml (if it even needs fixing, which is not a given), although first I'd have to understand a little of what all those seven files do 😛

2023-08-30 23:14:44

- reply

@npub1tx3…kxnn For the record, I appreciate that it has a pyproject.toml file with console script definitions and optional dependencies and everything, so people can just run `pipx run offpunk` (or `offpunk[http]` or whatever) if they like 😀 👏

2023-08-28 01:37:56

Octaman's mouth always hanging open 😮 and he still manages to be one of the most expressive characters
#Monsterdon

2023-08-21 00:37:56

- reply

@npub1gaj…78dz IMO definitely not something you should be expected to apologize for!