| 03:29 |
diana_coman |
whaack - re block explorer - do you plan to run this as a service as a free-for-all-to-push-whatever or as what? if after all those years and all the [http://ossasepia.com/2020/03/27/a-review-of-the-bitcoin-category-on-trilemacom/][experience made available, illustrated, brought home with a bigger hammer and pictures and |
| 03:29 |
diana_coman |
everything one can possibly think of], the moment a new bitcoin-related service is supposedly designed the "decision" ends up in short order "let's make a web interface and do it for good measure with whatever seems convenient what-can-possibly-go-wrong-or-why-would-it-matter", there's really not much point in pretending any "concerned about making sure" anything, seriously. |
| 03:30 |
diana_coman |
whaack - do me this favour and walk me through that decision process that ended up with "flask is the best option for this job", will you? |
| |
~ 10 hours 2 minutes ~ |
| 13:33 |
whaack |
diana_coman: What do you mean by a "free-for-all-to-push-whatever"? As for the flask decision, I guess it was a bit of a man-alone "this is what i think is the best decision based on my current knowledge." |
| |
~ 16 minutes ~ |
| 13:49 |
whaack |
And perhaps this should not even be a web service at all, I'm not sure. One could argue that someone w/ a local blockchain is only going to want to explore blocks they can see on their own node, and one without shouldn't be given this service for free (as it only encourages them to not run their own node.) |
| 13:50 |
whaack |
I had these doubts on my mind and decided to just tangentially mention them on my blog, I should have brought them here as direction questions. |
| 13:56 |
trinque |
whaack: did you happen to read the #t threads on this? |
| 13:57 |
trinque |
I'll point out, the content of *blocks* never changes. |
| 13:58 |
whaack |
trinque: No I haven't, do you have a link handy? |
| 14:03 |
trinque |
http://logs.nosuchlabs.com/log/trilema/2019-09-06#1934296 << I can only find me mentioning it. |
| 14:03 |
snsabot |
(trilema) 2019-09-06 trinque: iirc at one point also a block explorer was contemplated that'd shit static html |
| |
~ 32 minutes ~ |
| 14:35 |
whaack |
trinque: alright that's a good option. For the "block" explorer portion of the block explorer that works great. I also have functions to lookup information about a transaction or an address. The result of those two functions changes as outputs get spent / new txns come in that reference an address, respectively. |
| 14:50 |
diana_coman |
http://ossasepia.com/2020/07/01/ossasepia-logs-for-Jul-2020/#1028076 - ok; why do you think it's the best decision? |
| 14:50 |
sonofawitch |
2020-07-19 19:33:04 (#ossasepia) whaack: diana_coman: What do you mean by a "free-for-all-to-push-whatever"? As for the flask decision, I guess it was a bit of a man-alone "this is what i think is the best decision based on my current knowledge." |
| 14:52 |
diana_coman |
http://ossasepia.com/2020/07/01/ossasepia-logs-for-Jul-2020/#1028078 - whaack, especially when in doubt, just speak up in here, it can't possibly hurt, can it? |
| 14:52 |
sonofawitch |
2020-07-19 19:50:26 (#ossasepia) whaack: I had these doubts on my mind and decided to just tangentially mention them on my blog, I should have brought them here as direction questions. |
| 14:54 |
whaack |
diana_coman: Because the amount of time getting setup with flask is pretty low, it does what I need, it's been used before successfully, and I don't know of any other python web framework (other than django) |
| 14:55 |
diana_coman |
what is it you need done there? |
| 14:57 |
whaack |
I need to redirect web requests pointed at different urls to different python functions while passsing in query parameter values as parameters to those functions. The functions need to be able to return a string that gets sent to the user as html. I need to make sure that whatever process is doing this can handle multiple simultaneous requests without getting locked up. |
| 14:58 |
diana_coman |
why are *web requests* the need ? |
| 15:01 |
whaack |
I imagine that the easy access + a point and click interface with addresses and txn hashes being links to view information about themselves will be seen as useful. |
| 15:01 |
diana_coman |
that doesn't sound much like a "need", does it |
| 15:01 |
diana_coman |
fwiw, point-and-click interface is not the best I can think of, no. |
| 15:02 |
whaack |
I'm very willing to challenge that assumption |
| 15:02 |
whaack |
(My own assumption that point and click interface is a good option, that is) |
| 15:03 |
diana_coman |
at any rate, even taking all you said above as a given, entirely missing from the consideration in there is the amount of complexity this "time getting setup with flask is pretty low" brings in. |
| 15:05 |
diana_coman |
whaack - as trinque pointed out, the blocks themselves ie the basic "what a block explorer even is" are not changing; basically what you have there is an instance of a wider class of problems namely data visualization |
| 15:07 |
diana_coman |
as such, what you are in fact looking for (and what you were groping for earlier perhaps without quite realising it when you were asking for "what would be useful") are queries and data-views that harness best the data in there for various purposes |
| 15:07 |
whaack |
correct. |
| 15:09 |
diana_coman |
you know, depending on how deep you are interested to go with this, you can make it anything between a simple listing of blocks at the plainest end and otherwise a full-blown data explorer-cum-aggregator-cum-visualizer, whatever you come up with, at the other end |
| 15:09 |
whaack |
I think taking it one step at a time towards the latter end is the right way to go forward. |
| 15:10 |
diana_coman |
the thing is that the simplest thing is static, while the exploration required to find out the best *useful* selection out of the ton of possibilities further is not something best done via...web interface |
| 15:12 |
diana_coman |
further and leaving all the above aside for a bit, I can't quite understand why "web interface" rather than...irc bot? can even plug it in with the wot relatively easy and so leverage further existing systems for access and so on; even if "web interface", I can't yet see at all the case for something terribly complex |
| 15:12 |
whaack |
Okay. The other option I see then is to simply make READ on my database public and provide a script for some basic queries. |
| 15:12 |
whaack |
ircbot works well too |
| 15:14 |
diana_coman |
http://ossasepia.com/2020/07/01/ossasepia-logs-for-Jul-2020/#1028076 - this was specifically re the "push" command as I understood you want to provide this ie visitor to your website drops some raw transaction. |
| 15:14 |
sonofawitch |
2020-07-19 19:33:04 (#ossasepia) whaack: diana_coman: What do you mean by a "free-for-all-to-push-whatever"? As for the flask decision, I guess it was a bit of a man-alone "this is what i think is the best decision based on my current knowledge." |
| 15:15 |
whaack |
yes atm it is a free-for-all-to-push-whatever. perhaps this is a dangerous idea |
| 15:16 |
diana_coman |
whaack - you know, it depends also who exactly are your target users there; if you target those in dire need of "point and click", then I suppose you "must" web interface but in such case dunno, they already have a wide choice I should say, not exactly any lack of some website of this sort, is there? |
| 15:17 |
diana_coman |
otoh I don't know that people around here are all that much into gotta click something (not to mention that uhm, bitcoin-related even less, if at all possible, lol) |
| 15:18 |
whaack |
no there's not a lack of those websites, although they are all covered in ads and extraneous crap |
| 15:21 |
diana_coman |
heh, it's what the customers expect, what |
| 15:24 |
diana_coman |
http://ossasepia.com/2020/07/01/ossasepia-logs-for-Jul-2020/#1028110 - assuming you check what gets dumped in there, I don't see why would it be dangerous, no; only very wasteful as such and in my opinion a very poor strategic decision. |
| 15:24 |
sonofawitch |
2020-07-19 21:15:16 (#ossasepia) whaack: yes atm it is a free-for-all-to-push-whatever. perhaps this is a dangerous idea |
| 15:31 |
whaack |
diana_coman: With the ircbot interface, how do you imagine I should give the response to a query? As a post to paste.deedbot.org? The responses will be too long to be dumped into a channel. |
| 15:33 |
diana_coman |
whaack - probably a paste indeed |
| 15:35 |
diana_coman |
basically the answer on irc would be a link to the raw paste and all is fine and easy to get at. |
| 15:39 |
whaack |
diana_coman: Alright. Tbh I think the irc interface is a little more clunky, with one jumping around from client commands to pastes. |
| 15:42 |
diana_coman |
whaack - hm? being raw, you get the data with a curl/similar, no? |
| 15:44 |
diana_coman |
whaack - if you already have some sort of visualization that you know is great for whatever specific thing, sure, provide that in whatever form works best; but before you have that, there isn't much more than premature optimization that you can do there. |
| 15:44 |
diana_coman |
whaack - alternatively you can go the mpex route I guess. |
| 15:49 |
whaack |
diana_coman: the mpex route being you paste something signed somewhere and mpex reads it and responds with some signed message on a public board? or just that its a private service you connect to? |
| 15:49 |
* |
whaack bbl, ~20 mins |
| 15:55 |
diana_coman |
whaack - whether signed or not is up to you (ie 2nd consideration); the benefit of irc is that you can leverage existing voice and so on to *not* ask people to sign each and every query (because ugh, it is excessive really); but if you want as alternative non-irc, you can still do that quite easily with rather basic html. |
| 15:55 |
* |
diana_coman will be back tomorrow. |