Nice. This is a doc I downloaded from the web (public resource). DocDrop allows me to upload it by dragging and dropping, and start annotating as usual. This complementary functionality of Hypothesis makes social annotation complete.

DocDrop 的特點是，允許上傳某一文件，存於雲端。每個使用者有多少空間？能存續多久？這些都是未知數。

-- 註記實例：

翻譯體 translationese

Docdrop is another tool enabled by Hypothesis (Hypothes.is). Here's I'm annotating a portion of transcript of a YouTube video clip.

https://docdrop.org/

Can be used to create optical character recognition on .pdf documents and return documents with selectable/machine readable text.

https://web.hypothes.is/blog/exciting-new-feature-annotate-youtube-videos-with-hypothesis/

Wishing this was easier within YouTube directly instead of hidden within the LMS. Of course, there's always still https://docdrop.org/ for this instead.

Try also docdrop.

Or Hyperaudio (hyperaud.io, soon-to-be hyper.audio)

If I understand it correctly, with hyperaudio, one can also create transcription to somebody else's video or audio when embedded.

In that case, if you add to hyperaudio the annotation capablity of hypothes.is or docdrop, the vision outlined in the article on Global Knowledge Graph is already a reality.

Is it just me or is there something quirky with the rel="canonical" links for docdrop so that there isn't URL parity for the annotations on docdrop and the equivalent youtube videos. I thought it used to work, but when I visit the youtube version, there are no annotations on the page. I thought this used to work the way one would expect. What changed?

I'm noticing that in this particular case there are two rel-canonical links when one might expect only one:

<link rel="canonical" href="https://www.youtube.com/watch?v=XgMh6iuFbT4"><br /> <link rel="canonical" href="http://docdrop.org/video/XgMh6iuFbT4/">

Maybe the docdrop link should be a rel="alternate" instead?

cc: @dwhly

Brandon,

Obsidian Annotator (https://github.com/elias-sundqvist/obsidian-annotator), which I'm presuming you're using, looks like it's working from within Obsidian instead of a web page and is very clever looking, but without some significant work, I don't think it's going to provide you with the results you're looking for. It sounds like you want an all-public chain of work and Obsidian Annotator currently defaults to an all-private chain.

From my brief perusal of what's going on, the plugin appears to be tied to a single Hypothes.is account (likely the developer's) which defaults all annotations to private (only you) and as a result, even if you had the permalink to the annotations you'd not be able to see them presented on the web as they're all private and you wouldn't have access to the account. You could try filing some issues on the related Github repository to see if the developer might add the ability to make public annotations using your own personal account, which I'm sure would require your personal API key for Hypothes.is to be put into the settings page for the plugin in Obsidian. Another issue I see is that it's taking Hypothesis tags and turning them into Obsidian tags, which is generally fine, but the developer isn't accounting for multi-word tags which is creating unintended tag errors along the way that will need to be manually fixed.

If you're open to an alternate method of annotating and doing so in public, I can recommend a workflow that will allow you to do what it sounds like you're attempting. It starts with annotating .pdf files (either on the web, or as local files in your browser) in public using your own Hypothes.is account. (Most of this also works with private annotations, but if you want them to appear on public versions of web-hosted .pdf files with the same fingerprints, you'll want them to be public so others can see/interact with them.) Next set up the Hypothesidian script described here: https://forum.obsidian.md/t/retrieve-annotations-for-hypothes-is-via-templater-plugin-hypothes-idian/17225. There are some useful hints further down the thread on that page, so read the whole thing. The Github repository for it is here: https://github.com/SilentVoid13/Templater/discussions/191 if you need it. I've documented a few modifications I've made to the built-in template to suit my particular needs and which might serve as a template if you find it useful: https://boffosocko.com/2021/07/08/hypothes-is-obsidian-hypothesidian-for-easier-note-taking-and-formatting/.

You can then use the functionality of Hypothesidian to pull in the annotations you want (by day, by document, only your annotations, all the annotations on a document, etc.) For .pdf files, you may require Jon Udell's facet tool https://jonudell.info/h/facet/ to search your personal account for the name of the file or one of the tags you used. When you find it, you can click on it and it will open a new browser window that contains the appropriate urn file "key" you'll need to put into Hypothesidian to grab the annotations from a particular .pdf file. It will be in the general form: urn:x-pdf:1234abcd5678efgh9101112ijkl13. I haven't found an easier means of pulling out the URN/fingerprint of pdf files, though others may have ideas.

When you pull in your annotations you can also get/find permalinks to the annotations on the web if you like. I usually hide mine in the footnotes of pages with the labels "annotation in situ" and "syndication links", a habit I've picked up from the IndieWeb community (https://indieweb.org/posts-elsewhere). You can see a sample of how this might be done at https://notes.boffosocko.com/where-are-the-empty-spaces-on-the-internet where I've been doing some small scale Hypothes.is/Obsidian/Web experiments. (I'm currently using Blot.im to get [[wikilinks]] to resolve.)

Another strong option you're probably looking for is to use "via" links (https://web.hypothes.is/blog/meetvia/) on the URLs for your pdf files so that people can automatically see the annotation layer. (This may require whitelisting on Hypothes.is' end depending on where the files are hosted; alternately https://docdrop.org/ may be useful here.) Then if you've annotated those publicly, they'll also be able to see them that way too.

Another side benefit of this method is that it doesn't require the Data View plugin for Obsidian to render your annotations within Obsidian which also means you'll have cleaner looking pages of annotations in your web published versions. (ie. none of the %% code blocks which don't render properly on the web)

As I notice you're using some scanned .pdf files which often don't have proper OCR and can make creating annotations with appropriate Hypothes.is anchors, you might also appreciate the functionality of docdrop for this as well.

Given your reliance on documents and the fact that you've annotated some in what looks like Adobe Acrobat or a similar .pdf program, you might additionally enjoy using Zotero with Obsidian, Zotfile, and mdnotes as outlined here: https://forum.obsidian.md/t/zotero-zotfile-mdnotes-obsidian-dataview-workflow/15536. It's relatively slick, but requires additional set up, reliance on more moving pieces, and isn't as nice an overall user interface in comparison to Hypothes.is. It also misses all of the potential useful social annotation you might get with Hypothes.is.

Hopefully this is all reasonably clear and helpful. I'd be interested in hearing about options from others who are using Hypothes.is in conjunction with Obsidian or other related note taking tools and publishing them to the web after-the-fact.

Best, Chris Aldrich

And now, DocDrop.org lets me link directly to this moment in the video: https://docdrop.org/video/RVlYZ00vWPo/#:~:text=yes%20we%20killed%20the%20past

Note: First use of DocDrop with WeebitOffcourse.

If you are reading this "in context", then the video is keyed to this very quote.

Annotation cannot capture punctuation, or the 'tone', of John Lester as he delivers a virtual world keynote at the 2017 VWBPE; but, at least DocDrop.org does allows a fine-grain moment-capture notion-linking...or whatever it is that I am doing atm.

Testing out the OCR functionality of docdrop.org.

I'm noticing that the pdf fingerprint of this text somehow matches that of other texts as there are a lot of non-related annotations on this page.

Is docdrop doing something squirrelly with the fingerprint @dwhly?