07:08 <freemint> good morning

07:08 <Regenaxer> Hi freemint

07:17 <freemint> Good Morning Regenaxer

08:21 <beneroth> Good morning freemint :)

08:24 <beneroth> freemint, server side client/mobile detection (mainly via UA) is frowned upon because it it was a source of much incompatibility and confusion during the early browser wars, mostly because servers usually had no sane fallbacks but just responded with an error when it detected a client it thought not to be compatible with (or didn't like, error "better use browser XY" was a thing for a while, nowadays its mostly just Google doing such bs)

08:24 <beneroth> freemint, https://webaim.org/blog/user-agent-string-history/

08:26 <beneroth> also fashion is to split front-end development (JS+HTML+CSS+framework-of-the-month+UX-but-actually-its-just-design-without-userfriendliness) and backend (whatever+5-layers+SQL)

08:26 <beneroth> pushing frontend back into backend is not fashionable, at best you use the same language for both (JS, node.js)

08:27 <beneroth> pil GUI-framework form.l goes also against this fashion, because it has no client/server (frontend/backend) split/separation for the programmer

08:29 <beneroth> I personally believe working with UA would be much better than to non-discriminatory make every client download megabytes of stuff not relevant for that client.

08:32 <beneroth> technically I'm not aware of any technical reasons speaking against using UA, just take care to have a sane fallback. Some responsive design stuff cannot be taken server-side, e.g. screen orientation on mobile devices, so continued use CSS media queries (and maybe JS, but better if possible without) is meaningful.

08:35 <beneroth> also Google might discriminate against you (bad ranking) if they figure out you're presenting "different" pages on the same URL (differentiated by UA). of course the goal must be same content on the same URL, but I don't know how clever google analysis is in respect to such things. maybe Google doesn't detect it at all, maybe a little technical difference with no consequence to the content would get you a punishment.

08:35 <beneroth> so if you care about SEO / google ranking, better be careful with experimentation on that front.

08:37 <beneroth> also UA-header values are a complete mess, mostly because of historic bullshit reasons + laziness to ever clean it up. see link I sent you above. funny read.

08:53 <freemint> thank you for the excurse

08:55 <freemint> beneroth do you know whether schema.org is usefull for anything other than seo? (afk having a lecture)

10:13 <beneroth> freemint, afaik not really. well, its about making web-crawling cheaper (I assume one of the highest costs for search engine vendors in general, but also other crawlers besides search engines). so maybe you can benefit when you write a crawler targeting websites which comply to schema.org

10:14 <beneroth> you can view it as improving the readability/availability of the web. or as externalising the cost of writing a tolerant crawler.

10:15 <beneroth> as 80-90% of web traffic is probably crawlers/bots, it is kinda consumer-oriented *haha-only-serious*

10:16 <beneroth> (according to my personal anecdata its mostly seo optimizers and competition-surveillance)

10:16 <beneroth> (besides search engines, I mean)

10:17 <beneroth> (this claim is not the result of careful statistic work, just my personal impression)

10:35 <beneroth> Regenaxer, I have a input string which normally matches one of multiple possible patterns (or parts of them) (else its invalid input). Whats more efficient: (cond) with multiple (match)es, or making a FSM using (state) and popping through the chopped input string?

10:36 <Regenaxer> I would think 'cond' with matches

10:36 <Regenaxer> less looping on Lisp level

10:36 <beneroth> hm, T

10:37 <beneroth> yeah I go with matches, looks nicer in the code

10:37 <Regenaxer> 'sub?' cannot be used?

10:37 <Regenaxer> yes, that too

10:38 <beneroth> I need the reverse stamp (so.. $stamp)

10:38 <beneroth> 2018-09-12T17:04:06-05:00

10:39 <Regenaxer> Should be easy?

10:39 <beneroth> yeah no problem

10:39 <beneroth> just wondering about optimisations :)

10:39 <beneroth> prematurely

10:40 <Regenaxer> :)

11:27 <freemint> beneroth a schema.org like model lends itself quiet good to the picolisp db. that got me thinking

11:30 <beneroth> the graph model is very suitable to many applications.

12:09 <freemint> T

20:43 <beneroth> note to self: increase use of (cond) instead of (if)/(ifn) for validation logic

20:51 <Regenaxer> T

20:51 <beneroth> (put S 'stamp (cons (date) (time)))

20:52 <beneroth> (; S stamp 1) vs. (car (; S stamp)) *g*

20:52 <Regenaxer> yes, I use 1 often

20:52 <beneroth> the property is a cons, not a list

20:52 <Regenaxer> ah

20:52 <beneroth> therefore (; S stamp 2) -> SEGFAULT :)

20:52 <Regenaxer> (; S stamp -1)

20:53 <Regenaxer> Sorry, I use it in +Bag usually

20:53 <beneroth> aaaah

20:53 <Regenaxer> there 1 is right

20:53 <beneroth> yeah similar context here, memory object instead of bag

20:53 <Regenaxer> yep

20:53 <beneroth> ha its even written in ref for (get)

20:54 <beneroth> thanks! I was unaware of the possibility to give a negative number to (get) :O

20:54 <beneroth> makes sense!

20:56 <Regenaxer> All car and cdr accesses can be done this way

20:57 <beneroth> yes, and consequently also in (;) and (:)

20:57 <beneroth> nice!

20:57 <beneroth> I just haven't used (get) to access cdr's yet :)

20:58 <beneroth> I just told my gf how brilliant you are :)

20:58 <beneroth> (once more)

20:58 <Regenaxer> oh, blush

20:58 <beneroth> greetings from gf :)

20:58 <Regenaxer> Not brilliant, just logical consequences

20:58 <Regenaxer> Greetings back! :)

20:59 <beneroth> she says thanks

20:59 <beneroth> well

20:59 <beneroth> reflecting logical consequences in language design is brilliant

21:00 <Regenaxer> ok

21:00 <beneroth> its often lacking *cough*

21:00 <beneroth> :P

21:00 <Regenaxer> yes, strangely

21:01 <Regenaxer> The chained 'get' is another typical interpreter example

21:01 <beneroth> aye

21:01 <mtsd> Good evening :)

21:01 <beneroth> hm

21:01 <beneroth> T

21:01 <Regenaxer> the args are interpreted, not good to compile

21:01 <Regenaxer> Good evening mtsd!

21:02 <beneroth> T, very good argumentativ example!

21:02 <Regenaxer> mtsd: Son is sleeping?

21:02 <beneroth> (get) is a very important construct in picolisp, very handy

21:02 <beneroth> Good evening mtsd :)

21:04 <mtsd> Sons sleeping, I am up Penting :)

21:04 <Regenaxer> Very good both! :)

21:04 <beneroth> nice!

21:04 <mtsd> yes

21:04 <beneroth> how is going for you, mtsd ? :)

21:04 <Regenaxer> mtsd, did you put all swedish characters?

21:05 <Regenaxer> I changed Penti after discussing with your friend Fredric

21:05 <Regenaxer> Fredrik Gustafsson

21:06 <Regenaxer> "put all" I meant "define all"

21:08 <mtsd> yes, I told him to discuss with you

21:08 <Regenaxer> good

21:09 <Regenaxer> So I added F-D to define Alt-Chars

21:09 <mtsd> Sorry again about me being slow

21:09 <Regenaxer> no hurry!

21:13 <mtsd> no shortcuts, just have to work on this

21:14 <Regenaxer> Otherwise you need to enter language-specific chars via C-space or F-N (numeric)

21:15 <beneroth> F-D to define Alt-Chars? so user-defined chars?

21:15 <Regenaxer> right

21:15 <Regenaxer> I defined A-L ☺ and A-T ☹

21:15 <Regenaxer> (Laughing and Traurig)

21:15 <Regenaxer> also ♥ with A-W

21:16 <Regenaxer> Before that I entered F-N 9786 for ☺

21:17 <Regenaxer> But perhaps we have in the standard already all that is needed for Swedish?

21:17 <Regenaxer> After adding A-B for å

21:17 <beneroth> very nice!

21:18 <beneroth> could it also be whole phrases? :P

21:18 <beneroth> or only single characters?

21:18 <Regenaxer> A-B for å was suggested by Fredrik

21:18 <Regenaxer> No, only single chars

21:18 <Regenaxer> an integer

21:18 <Regenaxer> Same as can be entered with F-N

21:18 <beneroth> btw. Regenaxer, this might be interesting for your firefighter app (dunno if it allows any interfacing) https://www.gotenna.com/pages/mesh

21:19 <beneroth> bluetooth-connected antenna to form a mesh network in areas with no mobile reception

21:19 <mtsd> I will check swedish a bit better

21:20 <Regenaxer> beneroth, I'm pushing Briar with my friends since a few months

21:20 <beneroth> Regenaxer, ofc it would probably make your distributed db architecture even more complex when peer-to-peer is added

21:20 <Regenaxer> It can also make a mesh without internet

21:20 <beneroth> ah

21:20 <beneroth> nice

21:20 <Regenaxer> It is based on Tor

21:20 <beneroth> ok, then Briar is probably better. the things this gotenna.com offers are probably easy to implement yourself

21:20 <beneroth> right

21:21 <Regenaxer> yes, would be interesting to build oneself

21:21 <Regenaxer> in PilBox

21:21 <Regenaxer> But I think it does not help for the fire app

21:21 <Regenaxer> They need to communicate most of all with the base

21:22 <Regenaxer> which is too far away usually

21:22 <beneroth> well chat messaging between people on the ground?

21:22 <beneroth> also GPS

21:22 <Regenaxer> Yes, I have it

21:22 <Regenaxer> messages only to/from base

21:22 <Regenaxer> But lot of other infos

21:23 <Regenaxer> who is with whom, which vehicles etc

21:23 <mtsd> Time for bed for me, nice talking to you

21:23 <beneroth> yeah I mean peer to peer between people on the ground, not with base / sync with base later.

21:23 <beneroth> mtsd, good night, sleep well :)

21:23 <Regenaxer> Yeah, I'm also in the process of boading bed :)

21:24 <Regenaxer> Good night all! :)

21:24 <beneroth> good night Regenaxer :)

21:24 <beneroth> thanks for your advices, as always :)

21:24 <beneroth> sweet dreams :)

21:25 <mtsd> Good night : )

21:29 <beneroth> Regenaxer, gotenna uses radio (like walkie-talkie), so should have a better range than Briar