It's 2023 and I Can't Find Me

[Bill_French]

A few days ago, I tried to help another Airtable enthusiast concerning a question about timelines. According to this community platform, I have written about  3250 posts since 2018, and despite my grandfatherly age, I recall one such post about Vega-Lite that would be perfect to help this customer.

When a search engine believes you've made a mistake crafting your query, such as a misspelling, it tries to help. In this case, it demonstrates a significant degree of cluelessness.

Did you mean: negative-lite ... Seriously? And what will this brilliant discovery engine tell me about "negative-lite"? Feel free to dive into that rabbit hole. We need to press on with a point to this post.

Looking through all 42 pages of results did not contain the post I was looking for, but it did give me a few laughs like this one. It's entirely irrelevant to the query.


The Khoros search engine has no idea that Vega-Lite is an entity, much less an Airtable extension. This says a great deal about the indexing architecture under the hood and exposes the fact that the platform is not able to extract entities at all. This likely extends to people, places, third-party product names, and even essential Airtable constructs, including references to other community posts.

I've written extensively about Airtable views and the benefits of virtualizing data without making copies of records. Yet, a search for my name and "views", which proudly locates more than five-thousand hits, thinks the most relevant article mentioned by me is a post about Airtable's billing system. The results are peppered with many references about billing problems.

And don't tell me that I need to use the "Bill_French" user name format or the advanced search features; it's worse if you do that. I'm generally very kind about criticisms of products and services, but advanced search in Khoros is a joke. One of the first Internet search systems (AltaVista, circa 1995) ran circles around Khoros search, which should reflect almost 25 years in advancements. Since then, search has evolved to a very high bar and its technology is largely freely available through open-source projects like Lucene.

Since the Khoros platform replaced the previous community software, finding stuff has been a slog. I've tried all manner of techniques to locate historical passages that serve as the communities' second brain. Pre-Khoros I actually experienced moments while searching in the community for Airtable answers that actually led me to posts made by myself. These were delightfully embarrassing moments that no longer occur.

What's the remedy?

Khoros is a marketing platform disguised as a blogging platform hoping to sweep up and address all aspects of a knowledge-sharing and conversations platform. Expecting it to become a great search engine is not in the cards; it doesn't possess the underlying architecture to deliver on basic findability expectations. Yet, findability is crucial in its attempt to act in all three capacities, especially in the capacity of a community resource.

The only pathway (for me) is to create an external index of my content; I want all 3250 of my posts and the context in which they exist. Perhaps @ChrisShernaman could explain how to get that index. Is there an export feature? Or a script that uses the Khoros API that would extract all of my content?

[itsmike · Accepted solution]

Hi @Bill_French, in agreement with much of your post - discoverability has taken a dive since the new community was launched. A lot of valuable knowledge now seems harder to find. Hoping it's fixable but fear the issue may be how the platform is designed.

On your proposed remedy: I've gone ahead and extracted your post history and saved it to an Airtable base. You can find it here: https://airtable.com/shr1yffEwuJi2YWP8.

Hope the format's useful. After removing duplicates the total count is 3002 - I can share the raw CSV files if you'd like.

Cheers
 

[Bill_French]

That’s a very generous gesture. Oddly, I became so dismayed with how difficult it is now to lean on past posts and content that I harvested it all as well. But I moved all of it to Mem.ai which has a pretty good full-text search engine and a powerful classification system that magically builds a relationship graph between content objects and people.

In light of the situation, I’ve decided to reverse the engagement polarity for a number of reasons. My public contributions to the Airtable Community will be made available as shared content objects in Mem, negating any dependency on the community platform while also retaining control of the knowledge base.

Thanks again for your effort; it’s a lot of work. But, on the chance that you came up with a cool extraction technique, you should sell it to the many long-standing contributors whose historical legacies have been made invisible.

one of the benefits of findability is the rapid reuse of previous posts showing approaches and code that have been ostensibly peer-reviewed. Airtable has essentially muffled and then Toto-tilled seven years of ideas, sharing, and solutions.

[itoldusoandso]

I HATE THIS FORUM.  SUCK BIG TIME.  I HATE COMING HERE BACK. PITA FORUM.

[Matt_Boudreau]

Bill, I've learned tons from ready your various posts in the past years! I like the ones you talked about Apps Scripts; I used these patterns various times to generate pretty cool documents with my Airtable data.  You are a top pro -> I will try to find your shared content in Mem. Cheers.

[Bill_French]

Thanks for the kind words! This might help for future conversations.