Update project information, and future plans

[fabiankessler]

Hello, and thanks for this library. I've been using it for a while now.

I've noticed that there were no more releases for over a year, the last one being from 2014-11-09. That's a problem for my software, and probably for many other users also, since a year is a long time with technology that moves so quickly.

Windows 10 is out, and Edge is not supported here in the release. That's certainly a reason for users to turn away from this library.

After a bit of reading up, I've learned the following:

• Uadetector is made in 2 parts. The software (uadetector-core) and the user agent data (uadetector-resources).
• The data used to come from another source, user-agent-string.info.
• The publisher, Jaroslav Mallat (username @mallat), has discontinued the free data and now offers it as a subscription at duger.com see https://udger.com/about-us
• It seems to me that after this, the author of the uadetector André Rouél (username @before) has lost a bit of interest.
• Users keep contributing information in the bugtracker. And @mallat incorporates them into his data base. But they are not available to regular uadetector users.
• User @HaraldWalker, who used to have his own library at https://github.com/HaraldWalker/user-agent-utils has declared it to be in maintenance mode, and shows some activity here.
• Another library is at http://pieroxy.net/user-agent/lib.html GitHub https://github.com/pieroxy/java-user-agent-detection but it doesn't have many users
• Another Java library is https://github.com/ua-parser/uap-java coming from https://github.com/tobie/ua-parser they use croudsourced data from http://www.browserscope.org/ the Java version isn't the most popular port, but the whole thing seems to be widely used. Has anyone looked into their data? Could we use this?

What is the plan for the future? This software is too good to just give up on it. Uadetector is an authority. Mozilla https://wiki.mozilla.org/Compatibility/UADetectionLibraries checks their user agents against it (and against Harald Walker's User Agent Utils).

My current thinking is: The data from 2014 is available. We could build our own changes database that contains all the new stuff and overrides/fixes.

If I don't get updates from uadetector, then that's probably what I would do: write a wrapper around it, first detect what uadetector can't. Just that it always works better as a community effort.

For the time being, can you update the project documentation? That is:

• http://uadetector.sourceforge.net/ says monthly updates. Instead it should mention that there are no more free data updates, and possibly mention the option to license the data.
• the same on the GitHub page.

[jugglingcats]

Good summary of the current situation. Just for completeness, there is another Java client for udger.com data, described here: https://github.com/adhar1985/DIUASparser. But it has not been updated for a while and is not published to the public Maven repos.

[HaraldWalker]

If your project depends on it, then build a new version yourself. Uadetector now supports the udger.com database. If you don't need the auto-update feature you can still use it for free.

If anyone wants an official release with maven deployment etc., feel free to fork this project and offer it.

Uadetector always depended on the 3rd party user-agent database. The idea of a community based database came up more than a year ago. Personally I don't think this has a chance, at least not without cooperate sponsorship or donations. Maintaining such a database is a lot of work. No wonder that the user-agent-string database went commercial.

I suggested to udger.com some kind of affiliate system or revenue sharing, so that projects like uadetector have a motivation to continue their work but did not get a response after that.

My own user-agent-utils library is as you wrote correctly in maintenance mode. That means it gets about 4 updates per year to support new browsers but will not see major feature changes. That library has a different purpose than uadetector in combination with the user-agent database. It can not and will not be a suitable replacement.

I contributed the fix to make uadetector compatible with the udger.com database so that production systems can continue using it without having to implement a different solution.

[godenji]

Any update here wrt to alternative solutions? Just came upon this thread, and like @fabiankessler, noticed 2014.10 as the last Maven release.

Technology does indeed move quickly. Will dig through above options, was just curious if anything else has come up in the interim.