Q6Brain
When I right click on an attachment in a thought, and choose Select Related Thoughts what exactly is the PB looking at?  At first I thought it was checking the index of the document for Thoughts already in The Brain, but when I updated and saved the document (attachment) there was no change...

Anyone?

Thanks.
John O'
Quote
Moe
John, Thank You for posting. The Select Related Thoughts feature selects all Thoughts that have content mentioned in the attachment.

Thanks,
Moe
Quote
Q6Brain
Thanks for your reply but that's not the result I get...

Have a look at the attached. While The Brain matched four values ok, it  missed the two in red.

Click image for larger version - Name: Selected_Thoughts_Letter.jpg, Views: 183, Size: 287.86 KB
John O'
Quote
Q6Brain
Can anyone else take a stab at this? This feature of PB is exactly the way my clients expect metadata to behave...so they were a tad disappointed (and I was surprised) that submitting a new version of the document did not pick up the additional index terms.

Thanks.

PS. They also pointed out in my example that PB maybe should have picked up the address as a Coordinate as well.
John O'
Quote
Moe
John, The select related Thoughts feature examines the content of the attachment and selects the thoughts that it thinks are most relevant to the content based on their names. It's not a simple search and what is and is not selected can not be guaranteed.

Thanks,
Moe
Quote
Q6Brain
Thanks Moe, but with all due respect to the Brain development team (especially regarding the EKP) perhaps it should  be (a simple search) and it should (be guaranteed).

If I had an index of a little document like the one I show, and I could match the index values (any which apply) to existing thoughts - that I could click on to get more information - don't you think that would be a rock solid feature? As I said, many of the customers I have showed this to loved this, but want it to go farther.


John O'
Quote
Q6Brain
Here is another way I look at it...

If I can do a search on "Sasquatch" and get a result set back that shows the Product and its occurence in the document, why would it not show up in a Related Thoughts method?


Click image for larger version - Name: Sasquatch_Ale.jpg, Views: 135, Size: 101.04 KB
John O'
Quote
Moe
John, Thank You for posting. The search algorithm works differently then that of the select related Thoughts feature.

Thanks,
Moe
Quote
Q6Brain
Ummm...I kind of gathered that, but in a purely data-driven Brain like mine that turns the File --- Tag thought architecture on its ear it should be the same.

Think about all the overhead you could save Brain EKP users if metadata, for example, was managed this way. In my POC Brain, I have 5.593 thoughts, all of which are Q6 classified vocabulary values. When I drop a document into the Plex, any values I choose as metadata have associations already attached just like in my Thank You Letter example. 

If I was in another part of the Brain, and I found a document whose related thoughts were not immediately visible, I can tunnel into any other part of the Brain in one click.
John O'
Quote
Underdog
I can confirm John's findings that this feature is simply not working as claimed to work or as the user might expect it to.

Problem reproduction:
  1. Create city thoughts BERLIN; LOS ANGELES; PARIS; SYDNEY
  2. Create site thoughts FRIEDENSENGEL; SUNSET BLVD.; EIFEL TOWER; SYDNEY OPERA HOUSE
  3. For each of the thoughts in (1) create a txt attachment with the words of the thoughts created in (2), respectively
  4. Rebuild the search index or database (just in case)
  5. Run "Select Related Thoughts"
Results: no hits! (see attachment)

With other non-test thoughts having attachments I get bizarre unpredictable results which are meaningless, particularly when thoughts have more than one attachment.
- Michael
TheBrain 9.0.263.0 / 8.0.2.1 / JRE 1.8.0_161 / Win7 Pro, v. 6.1.7601 SP1 Build 7601
Quote
Moe
Underdog, Thank You for posting. The Select Related Thoughts feature works differently than search - if it had to search for every term in the document, it would take a very very long time and would not be useful. It not intended to be a comprehensive list.  

Thanks,
Moe
Quote
Q6Brain
Hi Moe;

I am more than likely reading more into your last post than you intended, but instead of "search for every term in the document" maybe what SRT should do is run the individual terms through the document.  Sort of like putting search on autopilot and using the vocabulary to test the new doc for hits.
John O'
Quote

Add a Website Forum to your website.

Newsletter Signup  Newsletter        Visit TheBrain Blog   Blog       Follow us on Twitter   Twitter       Like Us on Facebook   Facebook         Circle Us on Google+  Google         Watch Us on Youtube  YouTube       

TheBrain Mind Map & Mindmapping Software     Download TheBrain Mind Mapping Software