-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi,
I've finally updated and committed the test data used in the test mentioned earlier. (probably still 'committing' @ 45 kBytes/s) Because the structure of the files is different from earlier test data I have added a sub-directory to the existing test-data : https://svn.apache.org/repos/asf/incubator/devicemap/trunk/data/test-data/src/main/resources/test-data/UserAgent The files are described here : https://wiki.apache.org/devicemap/esjr/Test%20Data [not the most user-friendly editor, specially not if you're a vanilla MarkDown man like me ;-) ] Test Results ============ Using a test data set of 100,000 user agent strings (flagged 7 in UserAgentDetail.txt) on an iCore 7, 4Gb RAM test machine, over 3 runs : Code | average milliseconds per ua-string - -------------------------------------------------------------- OpenDdr | 30.83 DeviceMapClient (VB.Net) | 0.07 Fastest 'commercial' | 0.18 Resources/Release ================= I think we ought to look into and agree on the creation (if so desired) of n-gram based resources as opposed to the current regex-based ones. I also think that given the test results, even using the regex-based resource files, we should consider a release. Looking forward to your response, esjr -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.8 (MingW32) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQEcBAEBAgAGBQJSlbDUAAoJEOxywXcFLKYc3P0H/072HWeuK63vUSUv46wTG1J2 YlG0P64KZPXtemV1YVyRTmDO6CI2C/3BDyTo58J/ZbN7WgJ/WFLbrurqa6/ikAGx +mmqR2p2dyEEFIS4mmzfV3vadn32TxXXE93BpYlfjuNeJsG31Bu+l5WpZTo02Lis mnI+8HlFT67EEF2Rwqy1iVMJu+TLkoZTBzIZoUpoG0JfKSrYwgP7wdAhg+sVV+GW PnKjaE9xBIwtO/JE0JJ+Eo8BHiW++dcBzmW12YaibuFmI1hXFb+r8lKgm5GvtBRL 4h5usE3M0P6gOOmhfW8DhHxbVjVoi8qOqSWBtGTvTd6D6NReSOplhJ15zt52nGI= =EaDd -----END PGP SIGNATURE-----
