Today, my subscription to @premium expired. The quality of the platform is so bad, I might just delete my account.
It cost me $396 to rejoin the platform, but not with my old account @TS_SCI_MAJIC12 with 71,000+ followers and a community that garnered over 33,000,000 views per month to an account that is shadow banned and unable to break 3.69M impressions in a calendar year? Fuck you @X.
@TS_SCI_MAJIC12 @X @threadreaderapp can you roll this up please.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
@_devJNS For the reader/writer components of the code behind I didn't use any of these databases for the database. I ended up building a predictable format of a filesystem structure and then relied on Go's concurrency of accessing resources and TTL caching therein. jfkfiles.info
@_devJNS Now, the index on the other hand, that's what's curious to me and as I build out the application further, I find that I will be using Go for it, definitely, but the indexing will be decentralized, client cached and an unknown actual capacity. Can't know the full network size.
@_devJNS Given that limitation, I figured that each instance itself would have to rely therein only on what the filesystem had on it, and what was actually valid and could be served, and then the caching layer is in memory on top. Granted, I am looking into using something for that.