Fasten your seatbelts; we are ready to ship

There was a discussion going on in beagle mailing list sometime ago where I made a comment that I dont think beagle is newbie ready i.e. plug-and-play yet. Beagle fans did not understand my comment and people replied why they think I am wrong. I would be glad to be proved wrong; but all of their arguments were how they were using beagle since version x.y, how beagle was shipped and enabled by default since last z releases of some distribution and how someone is able to install beagle in a large enterprise. Duh! None of these prove that beagle is newbie ready. All they do is show that beagle works and even I know that. 0/100.

I have the feeling that some of the beagle devs and followers live in the garden of Eden surrounded by a high wall of reality. Sometimes they should go out in the streets, check the bugzilla of other distros, go through user blogs (which mostly contain complaints about how beagle does not work and how to disable it), and visit some user forums where a lot of questions are how to disable beagle from starting at startup. These are laborious jobs and not pleasing. A lot of them contain flames and invalid reasons. But almost always they are started by someone who found beagle causing trouble.

Here are some links which can make your task easy:

• Ubuntu beagle bugs
• OpenSUSE beagle bugs
• Fedora Core beagle bugs
• update Gentoo beagle bugs
  

I sometimes make the rounds and all I see are I make a point of uninstalling beagle on all my machines and The first thing I did after ... was to uninstall beagle and now my machine is happy. Silly men, how can they not like the doggy!

Subversion arrives. Finally!

KDE uses SVN for their source code management. GNOME used to use CVS till yesterday, which means beagle too was managed with CVS. I do not know the technical details. but time a again we did face technical problems with CVS. I was mostly told that life would be easier with SVN and lo behold! GNOME has switched to Subversion as its SCM (actually, still switching).

The last time this was tried by the awesome GNOME guys, they later found a glitch and had to cancel the migration. As a result I lost a commit that I made within hours of SVN migration. This time I will play safe and watch it for a few days before committing anything. If everything works out, life should be easier. Joe already cleaned up quite a bit of the unused files and directories, renamed the Evo-mail backend correctly and updated the links et al. A New Year with a clean, new repo. Sweet.

PS: There is one downside though. Joe (and others too) would like to use the SVN commit messages for creating Changelog files during creating a tarball. Which basically means others cannot observe the Changelog file between releases to figure out what was changed (neither I nor Joe updated the Changelog while committing, so this is a lame excuse). The real trouble is now I cannot write any lame jokes in my commit messages. Life would be serious now. Boo hoo.

My time with the doggie

Today I read about the Ohloh project (http://ohloh.net) and added my favourite project beagle (http://beagle-project.org) to it. I was curious what it actually does.

It took them nearly 4 hours to download and analyze the source code. But it was worth the wait. It showed some interesting statistics, like 122,885 LOC codebase, 82 direct contributors (committing in CVS) and 13 of them in last 12 months.

Just for a light comparison, Firefox has a codebase of 157,207 LOC, Amarok has 169,288 LOC and (take this) PHP 6.0 has 599,805 LOC.

It was also amusing to see my share in the project: http://ohloh.net/projects/3826/contributors/21154

beagle 0.2.14

Joe announced the release of beagle 0.1.4 few hours ago. Its a fascinating and shining new release containing exciting new features, lots of memory/speed optimizations and many bug-fixes as well. Here are the major ones which are readily visible:

• Indexes tar, gzipped-tar, bzipped-tar, gzipped and bzipped files, in the filesystem as well as in email attachments. The results show you the exact file in the archive that matched the query.
  
• Do some smart tokenizing to allow matching 001234 to a query of 1234 and better matching of file names. No more missing files.
• Beagle can find and extract data itself using its dozen or more backends. But sometimes its better for other applications to send data to beagle for indexing. Beagle had the infrastructure to act as a search/indexing service provider. The release contains an example C code to show how to do that; its pretty simple actually. Obviously python can also be used.
• Some cool signal mechanism which help to figure out what file in being currently indexed and for how long. This will be helpful if you feel beagle is taking ages to index some file.
• Use Xdg autostart mechanism to auto-start beagle. KDE4 will also implement xdg autostart mechanism. One more step towards being DE agonistic.
• The indexing information now explicitly mentions if the initial indexing is in progress. Also clients now have the option of being notified when the initial indexing ends.
• Lots of memory fixes. bhale just mentioned in the irc channel holy crap, startup RSS for beagled is 15m... beagled is below nautilus in mem usage...  im not believing my eyes :) Thank you for your myth on how beagle is a bloatware.
• API and beage-search support to know the total number of documents that matched any query. Not the superficial imposed limit of 100 documents.

Go, get it!

License to hack-debug-release

I start with the disclaimer that I am not a lawyer. I am also not very careful in reading the EULAs (End User License Agreement - a short acronym of a long term usually describing an even longer gibberish english text). When I release my first piece code (was it JBabel or mGet ?), I put them in GPL (possibly v1). Of course I didnot know what was I doing - I was merely completing a formality during the fascinating experience of sharing one's program. With time, and during development of Beagle, I came to know about the various licenses. It was about this time that the GPLv3 debate started. Also, I started releasing verious kde-beagle softwares which required me to figure out the correct licenses for them.

Liceneses, I feel, serve two purpose, demand credit where its due and specify the amount of responsibility.

Authors of some projects just want to release code without claiming any credit. I did so too. Only to be pointed out by someone that I have to actually specify the terms for others to use my code. Since I dont want to be credit for the code, its natural not to held responsible for its damages either. That needs to be specified too. One can either cook up one's own license or just pick one from the sea of licenses (http://www.opensource.org/licenses/) . The license I found to do just that is the MIT license.

MIT (X11) License: Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, ...

The license is itself not copyrighted, so it can be modified if needed. A rather free-to-do license, except for this clause following the above text

... subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

Wikipedia tells me that the latter part is necessary for the copyright laws of US and other countries. The credit part is followed by the author responsibility part

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT...

Often, people use this clause to explicitly deny responsibility

Except as contained in this notice, the name(s) of the above copyright holders shall not be used in advertising or otherwise to promote the sale, use or other dealings in this Software without prior written authorization.

I havent seen projects where the last clause wasnt present, so I assume it is not a necessary clause. I didnot use it myself when I released yaBi under MIT license. The core parts of Beagle and Mono are released under this license.

My other project, kBeagleBar uses the fancy LGPLv2. I had to use LGPLv2 because I was using some source code from kde libraries. Note that, LGPL allowed me to use the library without any restriction, but I copied some code from a library source file - so I had to use LGPL. Similarly, I release kontrol as (distributed as a part of kbeaglebar) which uses the code from GPLed kerry, so its under GPL too. Also GPL and LGPL are copyrighted by FSF. But then again, I wouldnt dare to modify them myself.

I am looking forward to GPLv3. No, not because I have any particular reason to like it. Because I hope to read it fully, understand it and read the various articles and blogs about it to understand the issues of free software licensing. Its always good to know where my rights end and start.

Even after all this, it will take me some more years to actually read EULAs. They are frightfully convoluted to confuse my little grey cells.