Young Hands
ossasepiaeuloratrilemaspykedtrinquetrilema-hanbotagriculturalsupremacy
44m1d 20h 38m72d 19h 31m3h 18m19d 8h 48m1d 16h 40m6h 22m
Ossasepia



whaack: diana_coman: EOD Report: G: I focused on fixing misunderstandings while working on my outline. WTI: Keep notes on what, if anything, I learned during the writing process for each article. B: I've let a few comments go without a response. WTI: I'll respond to them tomorrow and make it a goal to always at the very least acknowledge a comment within 24 hours.
whaack: diana_coman: The outline I made today: http://paste.deedbot.org/?id=3uXl Originally I was going to make an article about learning the command line w/ an example of a command I created, but I decided to move that example command into a separate article
whaack: http://logs.ossasepia.com/log/ossasepia/2020-02-26#1019466 lol well it's easier for me to know whether I get along with him than to know whether he is a good mechanic
ossabot: Logged on 2020-02-26 03:56:14 diana_coman: whaack: re mechanic, the more important part is how good a mechanic he is, rather than how well you get along with him, lol.
diana_coman: BingoBoingo: it's better but you seem to lack/not have yet clear a high-level structure of what exactly you are trying to figure out so that you get some clarity and can even evaluate how much/what you did/still need to do, hm.
diana_coman: BingoBoingo: looking at what you have there, how about this high-level structure: 1. Readership growth 1.1 strategy/approach 1.2 venues of interest (online and offline) 1.3 tools & implementation 1.4 discarded (ie already known to not-work, whatever they are but it's useful to list stuff & reasons anyway, lest one forgets or ends up otherwise with "nothing to show")
diana_coman: for 2. Contributor growth, it seems to be ~same really, could work perfectly fine.
diana_coman: BingoBoingo: here's a quick skeleton made for the readership growth based on the above structure; look at it and see how& where your items fit + what further space it opens up for clear exploration; structure similarly (and/or see what is different if anything at this level) for contributors growth; for both you really need to add and explore more offline options too, for sure; but unless you ...
diana_coman: ... structure it somehow, there's no chance to either get some clarity as to what *options* there are or to hold your ground and be able to have a proper talk when asked "what did qntra do for growth", you know?
diana_coman: it's one thing to say "no new contributors" and it's another thing to say "no new contributors after trawling x blogs, inserting y comments, picking z fights in t months; still in the pipeline to explore gab and twitter etc"
diana_coman: also, seriously now, qntra is not working in a void so you should get known too as the active voice behind qntra if nothing else; the point is that it's unseemly that all idiots get to talk unchallenged and shamelessly promote all the shit everywhere while qntra stands by like a wallflower or what?
diana_coman: BingoBoingo: if needed, do keep in mind/plan as well so that you rely a bit more on thimbronion and nicoleci or what other contributors you already have so that you do minimum for now on the writing as such and focus on growth first and foremost.
diana_coman: lobbes: how does that link-archiving work, can I have it in here too or what does it require?
diana_coman: http://logs.ossasepia.com/log/ossasepia/2020-02-27#1019476 good ideas there overall, indeed.
ossabot: Logged on 2020-02-27 00:21:53 whaack: diana_coman: EOD Report: G: I focused on fixing misunderstandings while working on my outline. WTI: Keep notes on what, if anything, I learned during the writing process for each article. B: I've let a few comments go without a response. WTI: I'll respond to them tomorrow and make it a goal to always at the very least acknowledge a comment within 24 hours.
diana_coman: whaack: you need to figure out better your "why" there for sure, heh.
diana_coman: at least the outline is good for highlighting this - you don't really know/have/are used to extract good whys.
diana_coman: whaack: e.g. "Piping lets one switch back and fourth between languages" - this is not a why and it can't possibly be a why; piping is very *powerful* but NOT because it makes you...switch languages, wtf! how do you even reason you'd link that in as a why learn it?
diana_coman: whaack: also, the "how to learn" is so, SO poor!
diana_coman: could add there "and stare at the examples for 1 hour while intensely piping languages"!
diana_coman: whaack: so, the good news is that those outlines ARE useful; the unexpected news is that there's way more work to be done on them before they are useful *for writing*; so far they are useful for highlighting where you need to do more work & thinking really.
diana_coman: http://logs.ossasepia.com/log/ossasepia/2020-02-27#1019478 - that may very well be but it still doesn't do anything; you know, it's *exactly* for this sort of reason that "LOC" ended up as "measure of software", with all the idiotic results of it: because it is indeed easier to ...measure; basically measuring stuff that is easy to measure instead of stuff that is *useful* to measure - horrible approach.
ossabot: Logged on 2020-02-27 00:25:28 whaack: http://logs.ossasepia.com/log/ossasepia/2020-02-26#1019466 lol well it's easier for me to know whether I get along with him than to know whether he is a good mechanic
BingoBoingo: diana_coman: Thank you. This is a good bit of direction, and getting myself out into the shit thumping idiots is an important point.
BingoBoingo: As is pulling back from writing myself to push on the growth front.
diana_coman: BingoBoingo: yw; have a look and a think, put the meat on those bones and/or ask if there's something unclear, see if there's perhaps more /different to add, iterate and come back, just keep going and it will get moving.
BingoBoingo: Alright, will do. I'm expecting incoming news from thimbronion later today, so that's one thing I don't have to worry about.
diana_coman: BingoBoingo: for that matter, what's your editing/publishing schedule/plan anyway?
diana_coman: why it was never requested/asked for is mainly because it seemed you have it well under control so yeah, no need to ask; but now I wonder - do you have a clear plan there?
diana_coman notes she is NOT A JOURNALIST
BingoBoingo: diana_coman: So far I take pieces as they come in, because the GPG wrapping obscures whether the story is breaking until I open the it up and give it that first pass read. That does leave my attention vulnerable to interruption, so editing and publishing is probably going to need some planning, though it isn't in the most urgent need for it yet.
diana_coman: BingoBoingo: that's one good point but not even the meat of it, huh; nevertheless, as you say, it is indeed not yet burning so it shall have to wait for now.
diana_coman adds a note to get back to this when there's at least some plan in place for tending the more pressing fires; sigh.
diana_coman: jfw, dorion what places/apps/whatevers are popular nowadays with the btc-interested people you know? it might help BingoBoingo to add to his list.
BingoBoingo inclined to take some time Sunday to go to the old city with the camera and cover the changing of the Presidents down here. May be some lulzy derps protesting etc, and the last photostory on Qntra did get some decent drama when I threw it on one of the LATAM subreddits
diana_coman: why not.
BingoBoingo not seeing an answer to "Why not?" beyond the time it'll take, which doesn't seem to outweight the up to now rare opportunity for boots on the ground.
diana_coman: BingoBoingo: meant to ask : what/how do you weigh whether drama is decent or not?
ossabot: Logged on 2020-02-27 09:49:21 BingoBoingo: inclined to take some time Sunday to go to the old city with the camera and cover the changing of the Presidents down here. May be some lulzy derps protesting etc, and the last photostory on Qntra did get some decent drama when I threw it on one of the LATAM subreddits
BingoBoingo: diana_coman: I weight drama when it draws curiosity or gets a rise. In that case, I made a since deleted submission, but someone else carried the Qntra piece around reddit asking about the story and receiving mostly dense responses http://archive.is/AigEd
BingoBoingo: Decent drama is drama that gets other people to assist the spreading.
jfw: http://logs.ossasepia.com/log/ossasepia/2020-02-27#1019476 - reminds me I've slipped on comment acknowledgement too.
ossabot: Logged on 2020-02-27 00:21:53 whaack: diana_coman: EOD Report: G: I focused on fixing misunderstandings while working on my outline. WTI: Keep notes on what, if anything, I learned during the writing process for each article. B: I've let a few comments go without a response. WTI: I'll respond to them tomorrow and make it a goal to always at the very least acknowledge a comment within 24 hours.
diana_coman: BingoBoingo: well, that sounds like some publicity so not-bad but it also sounds like an opportunity that could have been perhaps pursued for more, really; for one thing that someone who carried the piece around - did you talk to them? for the other - what/how many people/or-anything-visible/measurable did it bring back to qntra?
BingoBoingo: diana_coman: I'll have to follow these up more in the future. In this case all I did on that front was answer derps that yes, the Navy was involved, etc. That piece is recent enough that when access.log gets fed to AWstats, that there will be data on what that story drew.
diana_coman: BingoBoingo: you know, the more I think about it, the more I start doubting that the lack of a proper editorial plan is NOT pressing, huh.
ossabot: Logged on 2020-02-27 09:43:28 diana_coman: adds a note to get back to this when there's at least some plan in place for tending the more pressing fires; sigh.
BingoBoingo: diana_coman: This is starting to come together in my head as well. If we want decent drama, it's going to take some directed effort looking for lulz that connect.
diana_coman: BingoBoingo: iirc in the beginning when you started with cazalla you bought some journalism book(s) - were those on writing only or more and have you mastered it/them, did you stop there, have you read anything further?
diana_coman: BingoBoingo: and not *only* that, but seriously, you can't have a decent paper with ...whatever happens to fall from the sky or what?
diana_coman: and moreover, if it was indeed the case that just reading every day the news still was enough to get articles out with reasonable frequency, then it's probably *so much* that you've been leaving on the table that it ...hurts.
diana_coman: BingoBoingo: basically get it into your head that Qntra is a. a *business* not a hobby b. your future
diana_coman: so you know, it can't be just like that whatever happens to happen or something.
BingoBoingo: Beyond the AP style guide, no other books that I recall or that made the trip here. There indeed is quite a bit that's been left on the table over the years. Breaking the "whatever happens to happen" approach to things is something I've got to break.
diana_coman: for one thing, reading the about page, qntra is of course meant to be a proper paper aka reporting on everything of interest, not *just* on idiots and lulz, wtf.
diana_coman: so I don't know - was for instance the clean dissolution of pizarro reported? of nosuchlabs?
diana_coman: why not?
diana_coman: the start of tmsr-os project?
diana_coman: gah, I am not a journalist and I shall not become one overnight either but you can surely do better than that.
diana_coman: BingoBoingo: your overall goal with Qntra is to *own* the newspace in *everything* bitcoin-related; does this make full sense to you?
BingoBoingo: Alright. Expanding the coverage to everything is going into the plan, and does add some emphasis on the need to grow the contributor pool.
BingoBoingo: diana_coman: It makes sense.
diana_coman: well yes, of course you need more contributors, certainly.
diana_coman: BingoBoingo: the main question would be I supposed - do you actually want to own that space?
ossabot: Logged on 2020-02-27 15:50:26 diana_coman: BingoBoingo: your overall goal with Qntra is to *own* the newspace in *everything* bitcoin-related; does this make full sense to you?
diana_coman: suppose*
BingoBoingo: But it's become very clear the post-2017 "keep alive" slump and every assumption I've held since needs challenged.
BingoBoingo: diana_coman: I do want to own that space.
diana_coman: good then.
diana_coman: *that* is the long-term aim.
diana_coman: talking of short-term and long-term plans; *that*.
diana_coman: at this rate I'll end up on qntra's board without even realising it, ugh.
BingoBoingo: Thank you. I'm assembling the next revision, and I'll include editorial.
diana_coman: jfw: how's the fire-setting/fighting/tending?
BingoBoingo: diana_coman: That is a distinct risk you are running right now.
jfw: diana_coman: mostly cooled now and I'm working on some catchup.
diana_coman: ketchup!
jfw: One of the things was preparing a placement test for this prospective client; it was pointed out to us that, um, we can't just take their word that they know CLI stuff and sell them on one thing, then have it turn out that they don't and we need to spend a bunch of unexpected time
diana_coman: jfw: myeah, especially with words about what people claim to "know".
whaack: diana_coman: I learned now that one of my two fleets crashed about 9 hours ago. It seems to have been from some type of out of memory error http://paste.deedbot.org/?id=x6Mh I imagine that there is no point of standing up the fleet again until I've debugged the issue.
whaack: The dead fleet logged 850,000 lines before crashing. The one that is still going has logged over 2 million
jfw: We thought we might get that done yesterday but the meeting didn't get confirmed and is now tentatively Friday.
diana_coman: whaack: this memory issue thing is both ugly and really going on for way longer than it should; what's the root cause there anyway? the "great lisp environment" or your specific code or setup or what?
diana_coman: jfw: the joys of working with "busy" people; that's how it tends to go more often than not, indeed.
jfw: aye.
whaack: This is a different memory issue than before.
jfw: A bunch of dorion's time and some of mine was lost yesterday also on someone who "meet at my office" then turned out to be hanging out on the other side of town...
whaack: I do not know the cause of this one. But I did see a similar error when I had a memory leak in my program (I was previously storing some irc messages in memory, so eventually the program would run out of space)
diana_coman: jfw: at some point you do need to put in place some ways to dealing with this sort of people that mess you around though because it's usually not just a "one-off" either.
jfw: having more options is the best way I know to get more leverage on that kind of thing; but yeah, I could see some policy / sharper teeth helping anyway, and acting as a signal of our own value too.
diana_coman: jfw: it's not about "more options" as such, no, it can't be; first of all because if they are the sort of time-wasters, they will *anyway* waste *more of your time* than they are worth as they'll just keep dragging you down essentially; I get the trouble of "but it's anyway hard like hell to find even *that*" - it still doesn't mean it's fine to go along with it in that shape; hm.
jfw: it occurs to me, if we don't push back on time-wasting, we deprive someone of the chance to demonstrate they're capable of improving on it
diana_coman: jfw: exactly; and moreover, you give a signal that you do NOT want to give, there is that too and it's not unimportant.
diana_coman: regardless of how many/if you have prospects, you should still value your time properly because if *you* don't, then *they will surely not* and it's only going to get worse.
jfw: That signal being that we can be pushed around?
diana_coman: jfw: the way you described it so far, it would seem that is the signal you give, yes; it's fine to give the signal that you are acommodating but it should *have a cost* if they mess you about.
diana_coman: a cost for them, obviously.
jfw: right.
diana_coman: jfw: you know, it's amazing how quickly people "find time" to stick to what they said if they paid some advance or something; this is not to say it has to be an advance - find some way that works, but you can't just not do anything about it.
jfw: yeah, the incentive would depend on the situation but in general some way of recognizing the time value of ... well, simply the value of time.
diana_coman: jfw: and moreover of *your* time, specifically.
whaack: diana_coman: http://logs.ossasepia.com/log/ossasepia/2020-02-27#1019491 I can do better. To clarify the point about piping, would "One should learn the command line because the commands provide a means to write powerful and terse programs." be a proper "why learn the command line" and then a proper subcategory for that may be "One can create powerful programs because the output of command line
ossabot: Logged on 2020-02-27 09:15:17 diana_coman: whaack: you need to figure out better your "why" there for sure, heh.
whaack: tools can be wired to the input of other command line tools forming arbitrarily long chains via piping"
jfw: diana_coman: heh, right.
diana_coman: as you are there to educate them because you *know* something they *need* and they can't find anywhere else, so you're not just some replaceable resource they can't find anywhere
diana_coman: but all of this is several layers deep and wide and all that; and moreover I can't even say I fully get exactly how you positioned yourselves there so far and how you are working exactly, kind of hard to say like this.
diana_coman: it does come back to that don't sell yourselves short that I think I already repeated/pointed out in specific instances but you two know exactly what and how your are doing there.
diana_coman: and yeah, it's you there, not me.
diana_coman: whaack: that already sounds better indeed; go into more detail and explore the why(s) there as it will surely help you too, it's basically a way to better understand what it's all about anyway
diana_coman: whaack: the typical analogy is with lego pieces: basically command line tools offer small-and-clear-enough to grasp and use pieces that can be *further* combined together via piping to build larger and more powerful structures *gradually* (meaning while still controlling what is going on and without having to swallow huge complexity of the universal-tool-that-supposedly-does-it-all-nobody-knows-how)
diana_coman: jfw: fwiw, I know all about the "selling yourself short" through very direct and personal experience, lolz; only I was about 23-24 yo at the time and I tried to learn as fast as I could, lol.
jfw: diana_coman: I know missed or last-minute cancelled meetings or extended silences have happened a couple times at least with potential clients and partners, some more than others. Could be part of a selling selves short pattern, hm.
diana_coman: it's almost always a combination of factors, of course; but the part you control is still just one and it does matter.
jfw: I forget if I mentioned but it did feel like a novelty and some fun giving that stronger-on-sales presentation the other week.
diana_coman: whaack: another way to look at command line tools is that they are the only "reusable components" that actually work, despite all the wank and pretense in OO or whatever other similar camps
diana_coman: jfw: oh hey, you didn't mention it so far, no; and really glad to hear it, too! :)
jfw: diana_coman: thanks, and for the nudges at the time. (And silly of me to not bother responding there in real time.)
dorion: http://logs.ossasepia.com/log/ossasepia/2020-02-27#1019509 - whatsapp is ubiquitous in Panama and most of latam from my understanding. i gather telegram is popular in 'crypto' world. and signal is for people who want to imagine one phone app can be fundamentally better than another.
ossabot: Logged on 2020-02-27 09:44:50 diana_coman: jfw, dorion what places/apps/whatevers are popular nowadays with the btc-interested people you know? it might help BingoBoingo to add to his list.
jfw: ^, and facebook properties generally
diana_coman: well, BingoBoingo needs to trawl *all* the swamps, so *everything* goes in.
jfw: some others I've heard of are Slack and Gitter. I imagine trawling one swamp will naturally lead to others
dorion: for the telegram thing I've used tg, which is terminal proggy.
diana_coman: jfw: yw; I suppose the other thing I should point out explicitly -and not just for you really, for everyone- is that failing to value & acknowledge good things & people around explicitly and timely is both very easy to do and very damaging in the longterm (and no, even if it sounds self-serving, it's not this the reason I'm pointing it out).
diana_coman: jfw: ideally trawling one swamp should indeed naturally lead to the others, but given the poor record, I'd say it's best to have a list as long as possible to start with
diana_coman: BingoBoingo: ^
whaack: diana_coman: Okay. I will revise the outline again today and if time permits publish a short article explaining the command I made to turn a list of domains into a list of (ip, domain) for a host file.
diana_coman: whaack: all right; remember, your focus should be on doing things right first and foremost; the schedule/plan will have to adapt to *that*, not the other way around.
whaack: diana_coman: The point where piping lets one gradually build up a program from small, understandable chunks (that can be individually tested) is a good one that I had somehow missed
BingoBoingo: diana_coman: Indeed, I need all the swamps I can get.
jfw: diana_coman: right, it's damaging whether pointed out or not so I've appreciated that you highlight it.
whaack: diana_coman: and ack re adapting schedule to how long it takes me to do things properly.
jfw: whaack: I'm following from a distance here but 'ping' is definitely not the right tool for DNS lookup, heh. There are a couple options depending on which mess you prefer to depend on:
diana_coman: whaack: that can be individually *mastered* aka known in full detail, this is not about "tested"; imagine it like this: basically on one hand you can learn stuff in reasonably small chunks that are themselves useful and then as you learn more of them, the benefits increase not linearly but *combinatorially* because ....piping.
jfw: BIND, an old and ill-reputed monolithic DNS server, includes side utilities "host" and "nslookup". Glibc includes "getent hosts". I tend to use djbdns which includes "dnsip" and similar, though isn't widely installed by default.
jfw: whaack: ah, "dig" also comes with BIND I believe, though I see it as for more advanced tracing.
jfw: whaack: also the generating commands with awk and piping to bash strategy, while clever, strikes me as overkill when you could use a for or while loop
dorion: http://logs.ossasepia.com/log/ossasepia/2020-02-27#1019587 - what did you find were the most helpful changes ?
ossabot: Logged on 2020-02-27 16:38:18 diana_coman: jfw: fwiw, I know all about the "selling yourself short" through very direct and personal experience, lolz; only I was about 23-24 yo at the time and I tried to learn as fast as I could, lol.
dorion: I could see, "okay you missed a meeting, we'll reschedule with you once. if that's missed, it's $250/hr. ($250/hr is what I've seen from some lawyers for general consulting.)
diana_coman: BingoBoingo: btw, *still from that about* page, following the very link on qntra's creation, it turns out yet another clear aim & venues - did you ever/do you at all even have a list with the "competition" in the bitcoin arena? do you make sure they don't have "news" that qntra doesn't cover?
diana_coman: BingoBoingo: so do this properly and re-read *all* related to what qntra is even meant to be since apparently you never quite fully bothered to look or something; I can't begin to imagine exactly how but there it is.
diana_coman: BingoBoingo: extract from there everything of use, why wouldn't you even??
diana_coman takes a break to calm down already.
BingoBoingo: diana_coman: I do have such a list. Over time there was a tendency to retreat too much into the once fasionable bluster that "if things aren't happening in the Republic it isn't in Bitcoin", in retrospect it clearly wasn't the defensible line it may have seemed, and it is going in the discard pile for the editorial section of the plan which I'm hammering on atm. There's quite a bit of built up "old hands" shit of the sort that for
BingoBoingo: Qntra and other projects isolated, that has to be flushed.
BingoBoingo realizing through the revision of the plans that this is growing into a near, if not entirely complete operational reboot of Qntra.
BingoBoingo presents next draft of the plan http://paste.deedbot.org/?id=c8k8
whaack: jfw: http://logs.ossasepia.com/log/ossasepia/2020-02-27#1019613 Noted, I'll tryrewrite the command with a for loop. And all the commands you mentioned are indeed a more direct way to get the ip for a domain, "getent hosts" even returns (ip, domain) in the right format in one go.
ossabot: Logged on 2020-02-27 17:18:25 jfw: whaack: also the generating commands with awk and piping to bash strategy, while clever, strikes me as overkill when you could use a for or while loop
whaack: s/tryrewrite/rewrite
whaack: diana_coman: EOD Report: G: I took a stance against my one remaining drug addiction - caffeine. Instead today I drank some "organic indian chai tea." WTI: Try to not drink coffee for a week and see how that goes. B: Maybe related to the G, I was lethargic in the afternoon and took a nap that didn't seem to help. WTI: I'm going to go to bed very early tonight and try to stick to waking up at
whaack: sunrise (~5am here).

Random(ossasepia) | Download daily DB snapshot | Get Source Code