/tech/ - Technology

It would be far easier/cheaper to maintain chan aggregator make people write client against it than convince faggot like stephene to follow industry standard arbitrary one you invent at that

>>6664
>chan aggregator

explain, what would this be

bump,

would be interested in hearing what any of the professional devs here have to say.

Probably would extend the 4chan api as well

Why not use an existing protocol like NNTP (or nntpchan)?
Would imageboards need any additional server-side features?
According to >>564 it would be viable.

>>6839
I'm thinking more like a RESTful http/json based API.

Why not NNTP? in short, NNTP/usenet has a lot of problems which make it not suited to modern communication. NNTP/usenet and old school BBS's were fine in the 80s/early 90s but its way too dated at this point. Mainly, it isn't scalable because of its bloated replication architecture and the lack of authentication/identity/anti spam measures also made it susceptible to spam, and this actually ended up resulting in a few scandals with pirated materials and cheese pizza which is why ISPs mostly dropped usenet.

With the advent of the world wide web in the early 1990s old school networks like usenet slowly became the niche of a unix greybeards, to the modern internet consumer it's just obscure. Old school protocols like NNTP, IRC and even vanilla email are being replaced by more user friendly things like discord, reddit, etc. HTTP or even vanilla TCP/IP is a way more flexible foundation to build things on, even a decentralized architecture if thats what you want.

The web is already decentralized, no need to return to the 80s.

>>6665
filled with spam

>>6853
decentralization and federation sounds good until you realize 90% of other chans are fashy shit, mass shooter shit, incels, cheese pizza stans, autists, and other people that you would not want to interact with, with a ten foot pole let alone be in a federated network with, for both legal and optics reasons. Most of the relatively "normie" subject matter boards are only on 4chan mainly like /a/

>>6856
Nope, still sounds good. There's plenty of based lefty communities in the fediverse and I think the Trump brainworms have started to wear off on other 4chan alternatives. It's only scary to you because you don't have the spine to talk to those people and bring them back to your side.

>>6852
I really like the replication aspect, when it comes to text.
UTF-8 can be efficiently compressed, especially with LZMA.

I think modern browsers, that cache images of up to 320MiB, prove, image size is often overestimated.
If nntp clients had a whitelist policy in addition to killfiles, the volume would also be greatly reduced.

>lack of authentication/identity
There is RFC 4642, if you want something like that.
>anti spam measures
Is it actually worse than on imageboards?

What are other aspects of the nntp protocol you would consider bloated?

>>6918

I guess the question would be, given the disadvantages listed before, why not just use HTTP or thrift, grpc? The internet is already decentralized you could literally use any protocol. There's no reason to use an older protocol from the 80s when you could just do distributed/federated programming with any protocol, ex: fediverse just uses http/REST. Even plain binary TCP would work.

What exactly is the advantage of using NNTP when the rest of the world has moved on? There are also tons of libraries built and being actively maintained for those protocols, using NNTP would require doing greenfield development on that, pretty sure even Java EE for example doesn't have an NNTP library because people have lost interest in it a decade and a half ago at least.

I just want to know what the obsession is with using a protocol from the late 1980s long past the date when most languages/frameworks and even textbooks have stopped even caring about it?

It would be like If i was obsessed with creating an alternate internet based on the X.25 protocol because "TCP is, like, overrated man, they had it right the first time"

>>6922
I tried reading what NNTP was about, but quit because there is very little info on the topic. I assumed it was a protocol that was built on top of TCP, like an alternative version of zeronet.

Unless you're doing something highly specialized AND have institutional backing, there is no reason not to use HTTP over TCP.

>>6924
NNTP was an internet protocol used for usenet articles before the world wide web was invented in the early 90s. Its basically dead but theres one anon here who keeps autistically focusing on it, same as /g/entoofags talk about disabling javascript in browser because Stallman told them to like 10 years ago

>>6926
Vichan's code is littered with nntp protocol autism. It was never completed, afaik.

>>6922
My assumption was, that NNTP articles were close to a drop-in replacement for IB messages.
After reading the RFC, I must admit NNTP's retrieval of articles by traversing a GROUP is less than optimal for throughput and the structure of NNTP is inherently incompatible with the notion of threads, unless the server automatically moved articles to a new "subGROUP".

However I believe we should think about adopting some NNTP concepts such as:
board hierarchies (perhaps with their own overboards)
ability to retrieve resources through WILDMAT

Observing NNTP also gives us the chance to avoid its shortcomings.
Being able to retrieve messages by their ID is undoubtedly a good thing, but you could implement a request for threads or even boards.

Are operations on threads and boards (and/or matching against messages by date) sufficient to negate the cost of linear traversal, or are there more elaborate data structures, that curb the overhead of common usage.

>>6937
>Are operations on threads and boards (and/or matching against messages by date) sufficient to negate the cost of linear traversal, or are there more elaborate data structures, that curb the overhead of common usage.

Assuming the threads/posts are stored in a database, aren't most relational and even some non relational dbs based on B/B+ or modified B+ trees? If so average search complexity is O(log n), same for insert and delete.

Thinking about shit on the data structure level is a little too low level for web dev anyway since you would most likely be using a database/store someone else developed and its their job to make it efficient.

>board hierarchies (perhaps with their own overboards)
interesting, would the only point be for overboards?

>ability to retrieve resources through WILDMAT
is that some sort of regex search? Modern intelligent information retrieval techniques have probably obviated the need for that. Most likely using some sort of search engine thats already built although basic search functionality can be built pretty easily in a normal program.