lists of useful tags

[theo-armour]

@opentecture/owners

Do you have any help / thoughts / comments on how best to proceed with the following?

---

The current set of categories in 'Opentecture Bookmarks View-a-line' does not feel viable. If nothing else, a bookmark could and should be assigned multiple categories.

In essence assigning categories to bookmarks is a problem similar to deciding the length of a piece of string: the time it takes to decide is always longer than the piece of string. See also Clay Shirky's article (link below).

What we want is an infinitely adjustable who, what, when, where and why method we want to be able to answer questions like these:

Who is the creator of the page?

What is primary purpose of the organization

• Example three.js creates a software library

• What are secondary aspects
  • Example: Three.js produces: documentation, examples, issue management, utilities

So coming up soon there will be a 'Opentecture Bookmarks View-a-line-by-tags' script.

Current working set of tags

I have looked for lists of useful tags or keywords but have not yet found anything useful. Therefore I will start by building one from scratch as follows:

who: person, organization
when: time, day, year, life-cycle, short tern interest, long term interest, one-off/scheduled
what / produces: goods (tangible, physical) / services (intangible )
what / industry tags: academia, aec, furniture, industrial, software, hardware
what / physicality: online / downloadable / physical
what / media is: html, audio, video
what / cost: $$$ / fee / free / FOSS
where / location : country?
how / structure tags: formal/informal

More services tags with usage still to be determined:

• construction / fabricating
• content generating
  • portal
    examples: stackoverflow
  • articles - of longer duration / reference
    • examples: wiki, documentation, manifesto
  • news - timely
• consulting / design
• laws / standards / licenses
  • examples: webgl, creative commons
• code libraries
  • examples: three.js, python,
• management / logistics / QA / hr
• software
  • examples: Autodesk
• training / education / support
• utility
  • examples: AWS / ISP

Links

• https://en.wikipedia.org/wiki/Five_Ws
• http://www.shirky.com/writings/ontology_overrated.html

Organization

• https://en.wikipedia.org/wiki/Informal_organization

Products

• https://en.wikipedia.org/wiki/Intangible_asset
• https://en.wikipedia.org/wiki/Asset#Tangible_assets

Services
https://en.wikipedia.org/wiki/Service_(economics)

[afomi]

I have looked for lists of useful tags or keywords but have not yet found anything useful.

---

What this sounds like is a re-invention of a personal search engine with perhaps more internal data-classification features exposed to the user. This has quickly ballooned well-beyond Bookmarks to include new, notionally related information. What does "secondary information" mean? How does one use it? Calculate it? Maintain it?

What is the value? to whom?

---

The current set of categories in 'Opentecture Bookmarks View-a-line' does not feel viable.

What defines viable?

What we want is an infinitely adjustable who, what, when, where and why method we want to be able to answer questions like these:

infinitely adjustable seems untenable.

Who is the creator of the page? .. What is primary purpose of the organization? .. What are secondary aspects?

Again, what is the value? to whom?
Is it a blocker? If so, to what work?

[afomi]

a list of useful tags is called folksonomy
a list of mandatory tags is called a taxonomy or schema

In any case, if the goal is to manage and share bookmarks, I'd check out something like https://hypothes.is/. It supports tags, comments, page highlights, and has an API. I believe the bits are open-source too.

[schancel]

@theo-armour I have similar feelings to @afomi.

So far as I understand it you have a script which takes the JSON hierarchy from a web browser, and turns it into a convenient website for sharing. If you move to the tagging system you want, you'll need to maintain the JSON file by hand, rather than use the a web browsers UI.

Additionally, you'll lose the ability to quickly introduce a user to YOUR ontological framework for understanding how these websites relate. Providing an ontological framework is a powerful way to make user experience simpler and to share context. In contrast, if you move to a tagging system, you are asking people to discover things based on their own criteria -- if they have questions they want answered might Google be a sufficient resource?

It could be an interesting problem to solve, but implementing a query language and all the other accoutrements required will be quite a challenge.

[theo-armour]

@afomi @schancel

Have you read any of the several read me files? Have you run any of the demos?

For example:

From Opentecture Bookmarks Read Me

Problem to be solved

There are many excellent lists on the web. For example:

• https://frontendmasters.com/books/front-end-handbook/2018/tools/css.html
• https://github.com/sindresorhus/awesome
• https://github.com/asciimoo/ListOfGithubLists

And our very own list:

• https://github.com/opentecture/resources/tree/master/resources

You can spend many enjoyable hours exploring the links in these lists. There are several issues with this.

• Given hundreds of links, how can you prioritize what you want to see?
• Having seen a number of good pages, how can you come back to the list and return to the pages you like and not return to the pages you did not feel are useful to you?
• Given hundreds of links, how can you use search tools to help you find sites of current interest?
• Very long read me files with naked links are tedious and uninformative/uncurated
• Workflows for adding links to bookmarks in your browser then getting these links into markdown are complex

Objectives

• Build a list [of resources] good enough for inclusion here: https://github.com/sindresorhus/awesome
• Establish effective ways of managing curated lists
• Take a bunch of links from browser bookmark manager and get them into a JSON file quickly and easily
• View and manage links with a tree-view menu system

[theo-armour]

@afomi

In any case, if the goal is to manage and share bookmarks, I'd check out something like https://hypothes.is/.

If hypothes.is is a good way to manage a share bookmarks, why did you create https://github.com/opentecture/resources/tree/master/resources ?

[afomi]

why did you create https://github.com/opentecture/resources/tree/master/resources

As a lightweight way to share resources.

I did not intend it to be canonical.

[afomi]

@theo-armour - I did read through a couple of the README files (one in each directory). If there are others. I also had a chance to click around the page. See the bookmarks in an accordion tree. The bookmarks can have descriptions. Then a .json file can be exported. Then I assume it'd be used somewhere else (only in this library because additional custom data isn't supported by Chrome/Firefox)?

Regarding the following use cases:

You can spend many enjoyable hours exploring the links in these lists.

I acknowledge this is a Use Case, but spending many hours exploring these links does not seem to be uniquely valuable, or even something can can be accomplished with any amount of certainty.

Given hundreds of links, how can you prioritize what you want to see?

I wouldn't say that I prioritize hundreds of links. I don't use the internet that way. I browse and note, and if I need to come back to something, I put it in my own to-do list, which I currently use Trello for.

Having seen a number of good pages, how can you come back to the list and return to the pages you like and not return to the pages you did not feel are useful to you?

If the link is not useful, I'd delete it.

Given hundreds of links, how can you use search tools to help you find sites of current interest?

How is "sites of current interest" defined?

Very long read me files with naked links are tedious and uninformative/uncurated

Yet, there are multiple existing READMEs in this repo. And how does adding a layer of hand-rolled .html and .js make this easier?

Workflows for adding links to bookmarks in your browser then getting these links into markdown are complex

Is getting these links to Markdown a requirement? For whom?

---

As an exercise, bookmarks management via json lines is an interesting experiment.

But a world of static repos filled with compatible jsonlines files seems just as likely as a world filled with compatible XML files.

[theo-armour]

@afomi

What fun! You were able to be a "negative Ned" on all my thoughts. You exceeded my own "Doubting Thomas" predilections.

Unfortunately, doubt destroys disruption. And, reading between the lines I hear a "cease and desist" coming from you,

Yet, I have thousands of bookmarks - and I am not alone. It would be so nice if there were particular ways of sharing useful sets of bookmarks as parts of code libraries/reference works/lessons learned whatever. Not as parts of some giant amorphous web thing, but as a useful adjunct/curated to a GitHub project,

I do, however, hear you saying a "no" here. So what would you really like to see happening in the opentecture/mindmapping repo?

[afomi]

This repo was to explore connecting information in a 3 dimensional mindmap format. If managing thousands of bookmarks is valuable to you, by all means, continue with your work. The bookmark work is so specific it could and probably should be it's own project or proof of concept.

…

On Thu, Feb 21, 2019, 03:09 Theo Armour ***@***.*** wrote: @afomi <https://github.com/afomi> What fun! You were able to be a "negative <https://en.wiktionary.org/wiki/negative_Nancy> Ned <https://en.wikipedia.org/wiki/Debbie_Downer>" on all my thoughts. You exceeded my own "Doubting Thomas <https://en.wikipedia.org/wiki/Doubting_Thomas>" predilections. Unfortunately, doubt destroys disruption. And, reading between the lines I hear a "cease and desist" coming from you, Yet, I have thousands of bookmarks - and I am not alone. It would be so nice if there were particular ways of sharing useful sets of bookmarks as parts of code libraries/reference works/lessons learned whatever. Not as parts of some giant amorphous web thing, but as a useful adjunct/curated to a GitHub project, I do, however, hear you saying a "no" here. So what would you really like to see happening in the opentecture/mindmapping repo? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#62 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEJkFheYZFrUIvS7LOxaApuNoutVMO_ks5vPn5NgaJpZM4bCKEH> .

[afomi]

@theo-armour - if you're up to re-cast this bookmarks management feature within the context of the mindmapping project, that would be valuable - allowing every node to be edited similar to how a bookmark is edited.

a bookmark is one type of node/object. how is "listing useful tags" deliverable on a 3d mindmap? how are the tags persisted between users?

[afomi]

https://getpolarized.io/2019/03/01/polar-personal-knowledge-repository.html

[theo-armour]

My objectives ( this evenings update)

Make viewing many bookmarks pleasurable, memorable and speedy

• With tree in left menu / contents to right
  • Bookmarks may be viewed by browsing a tree of categories
  • Bookmarks may be searched for by typing in search text
  • Bookmarks may be located using tags

When multiple results are available bookmark targets are displayed with images and text in a Pinterest/Etsy/Fab store-like gallery sort of manner.

Clicking on one of the boxes, will bring up the full page for the site including user comments, date of latest site update, typical images from the site and more

Make curating many bookmarks easy-peasy

• With a click of a button grab bookmarks you have gathered using your browser and its bookmark manager
• Add images from the site and/or embed in the site in an iframe - if the target site offers that capability
• Tools are there to help you grab descriptions, text and images from the target site
• Add categories and tags is a matter of selecting items from drop-down lists
• Comments may be added and edited using Markdown or HTML

Make development of the viewing and curating tools simple enough for beginners to have fun with

• Hosting on no-charge static servers such as GitHub or Glitch
• Nothing to download. Nothing to install. All FOSS.
• Written in plain-vanilla JavaScript. Very few dependencies. Files of just a few hundred lines

---

I've been busy on the Spider things, but Add-a-line and her buddies are coming along gently as well