Re: Yet another browser extension for capturing notes - LinkRemark

[Max Nikulin <manikulin@gmail.com>]

Samuel,

since significant part of your message is dedicated to capturing of tab 
groups I should ask if you have tried version of LinkRemark add-on 
currently available from browser extension catalogues:

- https://addons.mozilla.org/firefox/addon/linkremark/
- https://chrome.google.com/webstore/detail/mgmcoaemjnaehlliifkgljdnbpedihoe

Groups of tabs or selected (highlighted) tabs are supported for 
Chromium, Firefox has no built-in tab groups, but it is still possible 
to capture selected tabs.

Your feature requests:
- Clean-up URLs. I have such idea, but I have not approached to 
implementation of it. Maybe URLs should be sent to another extension 
that excels in such task. If you have come comments which add-ons are 
great and which work rather poor, the suggestions my be helpful.
- Deduplicate URLs from tab groups. It requires some work to merge 
selected text, links, or nested frames from each tab. The complication 
is that some sites use internal navigation not reflected in location, so 
the same URL may have completely different content. Some sites have 
their top pages as canonical URLs, so some measures against false 
positives is required. Currently the extension may check if URL already 
present in org files. It requires https://github.com/maxnikulin/burl 
helper application that is in proof-of concept stage.
- Restore set of tabs. It requires some elisp code to iterate over 
subtree and to pick first "Link URL" or "URL" from description lists.

Currently I am thinking on some changes of interface since sometimes I 
just want to check if some URL is in my notes already. I would prefer to 
avoid adding more context menu items.

Additional details are inline.

On 17/01/2022 09:29, Samuel Wales wrote:
> On 12/26/20, Maxim Nikulin <manikulin@gmail.com> wrote:
>> On 26/12/2020, Samuel Wales wrote:
>>
>>> [... i can imagine great things possible with such extensions. for
>>> example, you could have sets of tabs, selected by right click in
>>> firefox, to save to a bunch of org entries.  then you could load that
>>> particular set of entries into firefox whenever you want.
> 
> interesting.  i do note tab selection features in recent firefox-esr
> and i was just assuming something like that.

There is no a ready to use recipe for loading saved tabs, but saving 
should work to some extent.

>>> You can do this with the "Copy all URLs" extension (ID:
>>> djdmadneanknadilpjiknlnanaolmbfk). Use this as the custom format (note
>>> the linebreak):
>>
>> I am almost sure that similar extension should exist for Firefox as well.
> 
> i think this is for copying all tabs, not selected ones.
...
> also i think this extension does not exist any more in firefox.

I have not tried them:
- https://github.com/piroor/copy-selected-tabs-to-clipboard/
- https://github.com/yorkxin/copy-as-markdown

>> - Are you going to capture reviews of "rice cookers" that could be
>> considered as ordinary pages or you are going to save items from online
>> stores?
...
>> Could you inspect head element of pages in
>> your favorite stores contains desired metadata using page source or
>> inspect element tools?
> 
> my web knowledge is too limited to understand your question, but i am
> just hoping it would capture ordinary amazon links, review sites, and
> so on.

It seems that quality of metadata in marketplaces like amazon severely 
depends on particular seller. The extension attempts to treat some data 
specially if there are microdata or JSON-LD with Product schema.org 
type. If I remember correctly, Amazon does not expose canonical link 
explicitly.

>>> [now if i can only debug the extra-blank-lines-in-capture problem.]
>>
>> Fully agree that it is really annoying. It is among high priority items
>> in my TODO list.
> 
> we might be talking about different thinks.  i am referring to
> something in org that adds blank lines when my particular org capture
> templates are used.

See info "(org) Template elements" 
https://orgmode.org/manual/Template-elements.html
:empty-lines, :empty-lines-after, :empty-lines-before
however I can not say that I really understand their meaning. Actually I 
do not mind to have empty line before next heading when refile is 
completed. My impression that it depends on number of empty lines at the 
end of capture buffer. I usually add some comments to captured pages.

On 18/01/2022 08:03, Samuel Wales wrote:
 > my amazon example was silly and confusing.  the point isn't shopping
 > for something; it's anything.  science papers, news outlets, nerd
 > blogs.

Scientific papers require more work, it is necessary to make them 
available to org-cite somehow. Some nerds use quite peculiar blog 
engines and strange setting of metadata. So shopping on some sites might 
work better than other cases.