📭 SpamSieve, worth it?
Let’s circle back to SpamSieve.
To be fair, the initial problem with SpamSieve not detecting spam messages was a configuration issue on my part. Simply put, I missed a step, which prevented the filtering of messages.
It’s now up and running correctly, tossing away 40+ messages a day with only two or three being false positives. And to be honest, those are kind of spammy, even though not actual spam.
After realizing my mistake, I deleted the Blocklist and started over. None of those rules were worth keeping as they were too specific. They would never catch another message. Checking for a sender of
firstname.lastname@example.org is pointless. The likelihood of that triggering again?
I still think that’s an issue with the way SpamSieve builds rules when you Train as Spam. They are too specific and will work once, then never again. If you keep that up, SpamSieve will be so choked with rules it will suffocate under its own weight.
A message will filter through 20k rules only for it to not match any criteria. Messages will take 10 minutes to get delivered and the entire system will come to a crawl.
While that seems like a total exaggeration, and nothing I’ve experienced yet, I’ve seen screenshots with 20k blocklist rules. That looks like a performance and maintenance nightmare. It also looks like the same incorrect usage I had.
Moving forward, if rules are needed, I will add them to the block instead of training them as spam. Or at least perform the training and edit the rule to be more general and work on more than a single message. Adding the message details to the corpus is a good idea, but a rule for every single email sender isn’t maintainable in the long run.
We’ll check on that after letting it run a while.
With SpamSieve now working, and trashing messages, in a good way, is it worth it?
It’s not an expensive app, nor is it a subscription, so that’s points in its favor.
The configuration isn’t difficult, but you need to be 100% accurate and double check your work or nothing gets filtered. That could be better.
The rules in the blocklist feel too specific. If you go that route, you’ll need to maintain them or you’ll have a mess on your hands.
The blocklist has a “hits” column, so in a couple of weeks/months, I can see which rules haven’t matched anything and remove them. Then again, maybe I won’t need to make any rules, and it’s a moot point.
It’s also interesting to note how many times the makers of SpamSeive (c-command.com) have added themselves and their products to the whitelist, 11 entries by my count. I see what you guys did there.
Right now, I’m torn over SpamSieve. It’s a decent app that’s blocking a lot of spam messages, just like it said it would. But, how much maintenance does it need to keep it from going off the rails? Hopefully none, and my first experience was just the result of my configuration missteps.
It’s time to let this run for a couple of weeks and see if it can redeem itself and run unattended without constant adult supervision.
Like I said, it’s trashing 40+ spam messages a day. The false positives dropped by the second day. I’m really hoping this takes me over the last hurdle and spam messages are deleted without me ever seeing them.