Cavallium

5 ms·

1.
by Cavallium 3y ago
A lot of vlang-related statistics are very suspicious, probably some of the metrics are boosted by a click farm or something similar. For example, 99% of the global google searches about "vlang" are from Beijing. https:/&#x
2.
by Cavallium 5y ago
I already explained to you why I didn't open more than one PR to the project, I'm not trolling anyone: without knowing Clojure the only contribution that I could do without spending literally days of work is just the use of System
3.
by Cavallium 5y ago
I never had the interest to "talk down Datalevin", I criticized the results because the title was quite misleading in the tones and the benchmarks were not rigorous enough to express conclusive statements.
4.
by Cavallium 5y ago
I opened a PR to use the more precise timers. This would not improve the performance of any of the tests, but it will improve their time accuracy. I highlighted some other problems related to the Lucene benchmark, but since I can't pro
5.
by Cavallium 5y ago
You completely misunderstood what this project does. This kind of libraries do not force you to return results in a specific way. Generally they allow the programmer to customize how the results are scored and chosen.
6.
by Cavallium 5y ago
That's not true. The real problem is that he is implying that his software is faster than Lucene showing data from a benchmark that has substantial flaws. The reason why the majority of criticism comes from Lucene users is because they
7.
by Cavallium 5y ago
I also suggest you to look at his "benchmarks" code
8.
by Cavallium 5y ago
He is beating lucene only if you look at his broken benchmark
9.
by Cavallium 5y ago
The funny thing is that the benchmarks that he wrote are more problematic than the T-WAND code itself. He didn't use any benchmarking library, and he used System.currentTimeMillis instead of the high precision timer available with Syst
10.
by Cavallium 5y ago
The benchmarks in question have several implementation issues, I reported them on GitHub. https://github.com/juji-io/datalevin/issues/created_by/caval...
11.
by Cavallium 5y ago
Now it's online
12.
by Cavallium 5y ago
Looking at gitly repository it almost has no features, only a fraction of the basic git features are present, and they are barely working. Why is it being advertised here on HN as a fast alternative of github/gitlab? P.S.: Probably the
13.
by Cavallium 5y ago
They also care about criticism https://imgur.com/a/8wYsj4C