Hi Schimon, can you tell use more about your end goal (end-user use-case)? Is it something for social sharing (e.g. I publish a book I'm reading, I want to allow comments, reactions, repeat, etc.) or is it something like having a collection of books/citation, etc? In the later case, you can have a look at XEP-0346: Form Discovery and Publishing, which is a way to share Data Form over Pubsub, I'm already using it to share various kind of things (TODO list, shopping list, tickets, merge requests), and I have plans for books too. It would actually great to have something usable with both, as both use cases are legitimate, in which case it could be describing the data to share and the fields to use in a Data Form, which could then be used either as attachment in a blog item (e.g. with XEP-0470), or directly with XEP-0346. Best, Goffi Le vendredi 7 juin 2024, 12:57:33 UTC+2 Schimon Jehudah a écrit :

Greeting Stephen! There is no special use case other than a dedicated node ID. The purpose is a a concensus node ID which should ease on interoperability betweeen systems. Suppose project Betula (links directory system) adds intergration with XMPP PubSub, and choses "betula" as node ID. https://codeberg.org/bouncepaw/betula/issues/9 BukuBot might expect "buku_public_links" and Movim might expect "movim_social_links" for bookmarks. An inconsistent node ID would hinder interoperability. Having a consensus node ID for this purpose would allow sharing of data without coordinating between projects, only referring to an XEP. Generally, I would suppose that it would be almost the same if not exactly the same as XEP-0472 in structure and data type. If you still think it is not needed, for the time being, please inform me and I will neglect this proposal. It is good to read from you. Best regards, Schimon On Fri, 7 Jun 2024 13:33:45 +0000 Stephen Paul Weber <singpolyma(a)singpolyma.net> wrote:

Good day! I retract from this argument due to XEP-0462: PubSub Type Filtering. Link: https://xmpp.org/extensions/xep-0462.html Thank you Stephen for raising my attention to pubsub#type attribute. On Fri, 7 Jun 2024 18:18:11 +0000 Stephen Paul Weber <singpolyma(a)singpolyma.net> wrote: Best regards, Schimon

Greetings, everyone! I am working on a "social" bookmarking system which is solely based on XMPP PubSub and Atom Syndication Format (i.e. XEP-0277 and XEP-0472). The benefits, as you already know* are: 1. The data is always stored on your account, because the system only provides an interface to make use of your own PubSub in a certain fashion, hence there is no need to worry whether or not the system shuts down or otherwise, so if someone decides to turn it off, someone else can run his own system and your bookmarks are still available. 2. The server does not store the data, so there is no need for a lot of storage to operate it. 3. No registration is required. Only an XMPP account. Note that for statistics and ease of intermediating people inside the system, it is essential to maintain a database, yet not all data has to be stored, which means that if the top 1,000,000 most popular links are important to you, for the sake of keeping the traffic, then there is no need for more than 200MB to 500MB of a storage disk for an SQLite database. * I am writing this information for people from outside of XMPP. Currently, the system is perfectly useable by a one man. Yet, I do need help in enhancing that system to allow access to multiple people at the same time. I was thinking of using the "cookies" system to store sessions and to link the received session value with the one which is store on runtime: accounts[jid] = XmppInstance(jid + '/blasta', password) sessions[jid] = str(random.random()) I am using Pyhton FastAPI and Slixmpp for that task. I would greately appreciate your assistance. Best regards, Schimon On Thu, 13 Jun 2024 14:26:44 +0300 Schimon Jehudah <sch(a)fedora.email> wrote:

Good day! I would also intend to make use of a couple of boolean fields. - Private - Read (mark a resource for "reading for later") * For option "private" I might use features provided by XEP-0060. * For option "read" and other custom fields, I might use The "atom:category" Element https://www.rfc-editor.org/rfc/rfc4287#section-4.2.2 So the required Atom fields are: - Title (plain text) - Link - Summary (plain text) - Category (tags) - Category (option read) And the required PubSub features: - Private (manage authorization) A note concerning to tags: Because I want compatibility with other bookmark managers, such as buku, then spaces are allowed, yet are not encouraged. Usage of colons and dots is encouraged instead; For example: brand:jabber country:ireland directory:recipe directory:tutorial license:bsd mr:william.cooper news:telecommunication niche:bookmarks service:isp service:restaurant service:vps store:beverage tutorial:cs tutorial:python tutorial:slixmpp utorial:xmpp etc. The question is: Is this proposed use case justifies a new Atom Over XMPP and PubSub (XEP-0060) based XEP, similar to XEP-0277 and XEP-0472? Please advise, Schimon On Fri, 19 Jul 2024 19:51:12 +0300 Schimon Jehudah <sch(a)fedora.email> wrote: > Pardon for posting this message here. > > The message has been forwarded to JDev and a new friend from Lithuania > has promptly helped! > > I think I want to raise the argument about a new XEP, again. > > Jérôme and Stephen, I would want to direct an issue to you. > > I am working on a bookmark sharing system based on PubSub and Atom > Syndication Format (i.e. XEP-0277 and XEP-0472). > > > I might add content (i.e. storing a whole article as Movim does) in > future, but I doubt I would do so soon. > > Yet I do not think I am interested in having HTML tags inside fields > "title" nor "summary", despite Atom Syndication Format allows > "type=xhtml" and "type=html" for each. > > Also, Libervia and Movim might be interested to interact with the > system that I am building, and they might determine that HTML tags are > fine. > > 1) Should I ask for a new XEP? > 2) Should I clean sanitize, so called, HTML tags from titles and > summaries? > > What should I do? > > Thank you, > Schimon > > On Thu, 18 Jul 2024 19:29:46 +0300 > Schimon Jehudah <sch(a)fedora.email> wrote: > > > Greetings, everyone! > > > > I am working on a "social" bookmarking system which is solely based > > on XMPP PubSub and Atom Syndication Format (i.e. XEP-0277 and > > XEP-0472). > > > > The benefits, as you already know* are: > > > > 1. The data is always stored on your account, because the system > > only provides an interface to make use of your own PubSub in a > > certain fashion, hence there is no need to worry whether or not the > > system shuts down or otherwise, so if someone decides to turn it > > off, someone else can run his own system and your bookmarks are > > still available. > > > > 2. The server does not store the data, so there is no need for a lot > > of storage to operate it. > > > > 3. No registration is required. Only an XMPP account. > > > > Note that for statistics and ease of intermediating people inside > > the system, it is essential to maintain a database, yet not all > > data has to be stored, which means that if the top 1,000,000 most > > popular links are important to you, for the sake of keeping the > > traffic, then there is no need for more than 200MB to 500MB of a > > storage disk for an SQLite database. > > > > * I am writing this information for people from outside of XMPP. > > > > Currently, the system is perfectly useable by a one man. > > > > Yet, I do need help in enhancing that system to allow access to > > multiple people at the same time. > > > > I was thinking of using the "cookies" system to store sessions and > > to link the received session value with the one which is store on > > runtime: > > > > accounts[jid] = XmppInstance(jid + '/blasta', password) > > sessions[jid] = str(random.random()) > > > > I am using Pyhton FastAPI and Slixmpp for that task. > > > > I would greately appreciate your assistance. > > > > Best regards, > > Schimon > > > > On Thu, 13 Jun 2024 14:26:44 +0300 > > Schimon Jehudah <sch(a)fedora.email> wrote: > > > > > Good day Jérôme! > > > > > > Thank you for joining to this thread and sharing insights! > > > Please pardon me for not posting to your post ID. > > > Your message was not delivered to my email. > > > > > > > Hi Schimon, > > > > > > > > can you tell use more about your end goal (end-user use-case)? > > > > > > > > Is it something for social sharing (e.g. I publish a book I'm > > > > reading, I want to allow comments, reactions, repeat, etc.) or > > > > is it something like having a collection of books/citation, etc? > > > > > > > > > > It is more similar to the latter: Storing of bibliographic > > > references on PubSub nodes and also publishing and sharing > > > (restricted to authorized contacts) of those references. > > > > > > Bibliography on PubSub: > > > * Storing (privately); > > > * Publishing (publicly); > > > * Sharing (restrictedly). > > > > > > I am currently focued on systems which are commonly referred as > > > "bookmark managers" or "links directories". > > > > > > Example: https://codeberg.org/bouncepaw/betula/issues/9 > > > > > > > In the later case, you can have a look at XEP-0346: Form > > > > Discovery and Publishing, which is a way to share Data Form > > > > over Pubsub, I'm already using it to share various kind of > > > > things (TODO list, shopping list, tickets, merge requests), and > > > > I have plans for books too. > > > > > > > > It would actually great to have something usable with both, as > > > > both use cases are legitimate, in which case it could be > > > > describing the data to share and the fields to use in a Data > > > > Form, which could then be used either as attachment in a blog > > > > item (e.g. with XEP-0470), or directly with XEP-0346. > > > > > > > > > > This appears to be a wonderful idea! > > > > > > I will strive to make BukuBot utilizable with PubSub in both > > > fashions you have described. > > > > > > > Best, > > > > Goffi > > _______________________________________________ > > Standards mailing list -- standards(a)xmpp.org > > To unsubscribe send an email to standards-leave(a)xmpp.org > _______________________________________________ > Standards mailing list -- standards(a)xmpp.org > To unsubscribe send an email to standards-leave(a)xmpp.org

Good evening, to one and all! The proposed bookmarks system has been published and it works as follows: Nodes are referred to as directories/folders and are managed by Access Models. Public items xmpp:bibliography:0 (Open, Presence, Roster) Private items xmpp:bibliography:private:0 (Authorize, Whitelist) Read items xmpp:bibliography:read:0 (Authorize, Whitelist) Nodes Read and Private are the same, though there might be some rare cases which Private be given access. Node Read is for convenience and is intended to facilitate management of data. The properties of Payload / XML / Atom: Title (text) Summary (text) Link Category * Plain text only. * No HTML allowed. We are considering to make use of Atom attribute "rel" for Link, in cases which targets are accessible in several forms or resources. Examples: * JMP https://blog.jmp.chat/ xmpp:blog.jmp.chat?;node=urn%3Axmpp%3Amicroblog%3A0 * Dig https://digdeeper.club/ http://dv2recs7spoc4vquwnt7gotljvot34i4am4xqsfslhfvdymgjfoa.b32.i2p/ https://digdeeper.her.st/ Yet, I sense that the base properties are complete. Please see screenshots https://git.xmpp-it.net/sch/Blasta/src/branch/main/screenshot The Blasta system, as a sharing platform, is still in testing, yet it seems to be fully operational, *and* the bookmarking practice per JID appears to be stable and valid. Shall we define an XEP for bookmarking system, similarly to urn:xmpp:microblog:0 of Libervia and Movim? Link: https://git.xmpp-it.net/sch/Blasta Kind regards, Schimon On Sun, 21 Jul 2024 13:43:41 +0300 Schimon Jehudah <sch(a)fedora.email> wrote: