I notice that team_rocket_qwerty uses Elasticsearch to do full text search and analyze fake reviews. It has machine learning capability so I think it can also be trained to detect patterns. It's possible to find similar reviews or fake reviews if he does it well.
.
Elastic Stack will provide much more powerful system for searching reviews if TER were to use this as their database for all reviews. Imagine searching for any particular phrases or words in all of TER reviews. We can search based on all current search fields in addition to General Descriptions, Juicy Details, and words in Appearance, Performance, Attitude, Atmosphere, and Session Location.
.
Who wouldn't want to search for "hot passionate big booty juicy 5'7" 24 years old D cup no implant athletic Asian in Los Angeles with great English and sweet attitude"?

Ha, it's flattering someone noticed. I dump reviews into both postgresql and elastic, yeah

Some of my favorite features of elastic in context of analyzing reviews are

morelikethis (mlt) :  
http://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-mlt-query.html

And
and significant term aggregation:

http://www.elastic.co/guide/en/elasticsearch/reference/current/search-aggregations-bucket-significantterms-aggregation.html

They are really, really good for finding similar reviews of same style.

And yeah, full text search is super powerful and should be on any review site imo.

.
I don't know ElasticSearch, but ... does it use a lot of TER bandwidth (if it is an external search app)? Keeping the searches simple and internal could save a lot of bandwidth and CPU.  
.
Some previous suggestions, going back several years now.  
http://www.theeroticreview.com/discussion-boards/suggestion-and-policy-4/how-about-a-general-or-flat-text-search-15330?frmSearch=1#15330
http://www.theeroticreview.com/discussion-boards/suggestion-and-policy-4/plagiarism-checker-for-ter-admin-to-check-reviews-14632?frmSearch=1#14632
http://www.theeroticreview.com/discussion-boards/suggestion-and-policy-4/re-i-highly-doubt-the-same-code-would-work-14635?frmSearch=1#14635
.
All reviews (General and Juicy) and our PMs are plain text. By pre-indexing the text, a simple search could very quickly, as in miniscule CPU time and mere seconds of clock time, find text strings. Because it's all plain text, it wouldn't take up a lot of space as pictures and other complex info would.  
.
Every college student knows about plagiarism checkers that their profs use to find cheaters.  Is any of that code available for free? Once again, the relevant TER content is all plain text: no fancy formatting.  
.
But I think that most of us are just interested in plain search and not plagiarism search.
.
It has also been suggested that TER could switch the Discussion Boards to open source Forum Software. Most Forum Software has built-in mail (like TER PM) and most of those are searchable.  And most Forum Software also has built-in Forum search tools that are better than TER's fairly simple and limited Search.
http://www.theeroticreview.com/discussion-boards/suggestion-and-policy-4/like-real-forum-software-16319?frmSearch=1#16319
.
http://en.wikipedia.org/wiki/Comparison_of_Internet_forum_software

Every time I’ve suggested a better search engine or forum software, even when it’s free to them, I’ve been told “we don’t have time or make enough money to afford it”.

But we can dream. And there are some great suggestions here.

I think that I've lived through 3 (or 4?) versions of TER Discussion Board ("Forum") software. I got the impression that it was always custom written, internally to TER. But that means months of writing and debugging code followed by MORE and MORE debugging after it's released to the TER-public and they start complaining.
.
The open source (free!) Forum softwares have large teams of developers and maintainers as well as large communities of volunteer bug checkers and helpers. TER could probably customize some free forum software with help from the global community to get a pretty good result. I can think of 4 levels of "membership": 1. Admin 2. VIP 3. Provider Free Limited VIP 4. Regular Schlubs. Pick the right software and get help to structure and maintain those levels (or more!).
.
One of the biggest problems with any sort of migration is porting the database or archive. Once again, I'd bet the "global community" could help TER programmers with that, too.  
.
NOTE: Forum software and REVIEW DATABASE are not the same thing! Forum Software is for the Discussion Boards. I think the Review Database is totally custom and development and maintenance has to remain internal to TER.

I've been pushing Open Source Forum Software and pre-indexed flat text searching on S&P for 5 or 6 years, at least. I can't find any of your posts on the topic. Are you plagiarizing me? HELP! Get a plagiarism checker ASAP!  

I realize I didn’t specify. But I’m talking about at other sites. And those were mostly emails directly to the site admin staff. Please accept my apology for causing confusion.

-- Modified on 6/2/2020 8:13:16 PM

What do you mean by bandwidth exactly?

They are storing reviews and profiles in some persistent storage on the backend, such as a regular relational database. They index certain fields in certain tables, just like you would normally with a database. They don't store any pics/binary data obviously, so basically it's database full of tables of string/text fields.  

Elastic is essentially another type of database. You index documents and store them. Then you retrieve them. Instead of 'schema', you have mappings. You have fields that have built in analyzers  

Full text search is possible in both. Is it 'expensive' in both? Yes. But if you know which field customers will search by, it's not that bad.

So, elastic is something you run on your backend, just like a regular database. It probably does take more resources than a compact db, but I don't think it's a big difference. Usually if you need to update fields often you'd choose a regular db, but if you mostly insert and retrieve youd use something like elastic or a kv store.  

Part of the reason elastic is good is exactly because of the built-in text analyzers for text fields. Elastic is often used for logging. Writing logs to files is super expensive and you need to manage those files and searching through huge files is painful. Searching through elastic is great and you'd usually index each day of logging so you can delete old indexes you don't need.  

The problem is, migration can be painful. Although, with only 2.5 million entries of reviews on TER that aren't updated often, I think it's quite manageable in 2020. However if they have a separate table for girl profiles and they have relationships, things start to get murky fast.

No problem. I don't think TER will be making any major changes in the near future. I am hoping for a better future, a stronger revival of TER, and then, MAYBE, improved Discussion Board and Profile / Review interfaces.