Criticism is one thing. Ranting and telling them to "get your shit together" is just uncalled for. If you have a problem or issue let them know so they can fix it.

Sadly they are missing any information about their testing procedure. How many runs did they do? How did they average the data? Etc.

"You dun goofed," he wrote. "Badly. Get your shit together, because a 20 per cent error rate for spam detection is making your spam filter useless."

I periodically go through my spam folder and so far I have a 0 percent error rate. I would bet the vast majority of gmail users are in the same boat. Unless he is paying for his gmail service this type of bitching isn't really called for.

Also, I am surprised he's not using the tagging feature. I would have to validate but I know at some point if an email was tagged it would bypass the spam filter. That would pretty much solve his problem...

A video with the data was on youtube for a short time but we eventually pulled. If anyone happens upon the data I would love to see it.

JASON

?? Assuming typo.

Anyway, I have been having some difficulty finding a use case for CouchDB. In the instances were I need a JSON document database I am not leaning towards ElasticSearch. No doubt there are situations were ElasticSearch as a primary datastore are not optimal but in the use cases I have currently I'm having a hard time justifying using CouchDB as the repository of truth and ElasticSerach as the indexing engine.

This is the primary reason I don't have any sort of smart watch. The size are simply too large to be usable with my smaller wrists. I'm looking at something like the vevosmart as an alternative.

I forked over the money to switch to "pro" and there are some interesting charts/data in there. Nothing that will make you faster or a better cyclist but just fun datamining. Hopefully he adds some custom reporting tools so I can develop my own charts.

Too many. The biggest chunk of time is cycling but coding, building, and art all factor in there too.

On your side note, I would suggest sticking to Python 2 as the majority of learning material is still built around that. Then after you have created some programs go back and convert them to Python 3. That should allow you to learn the basics and then learn the differences between the versions.