Technorati Tags: beagle, gnome, mono, ubuntu, gutsy, deb, bzr, bazaar, launchpad

With everything that has been swarming all over my plate lately, I haven’t had a chance to really keep on top of the Beagle packages in Ubuntu, and as a result, they are currently pretty crappy. I have a branch (meant to be feisty-updates, but I was in a hurry, and didn’t feel like branching), with a building deb configuration for Gutsy. I hope to have binaries/sources available for testing later this week.

The branch is hosted here:

https://code.launchpad.net/~kkubasik/beagle/feisty-update

Just do the following to try and build:

   1:  bzr branch http://bazaar.launchpad.net/~kkubasik/beagle/feisty-update

   2:  cd feisty-update

   3:  sudo apt-get build-dep beagle

   4:  bzr builddeb -w --split

A recent e-mail mentioning my Blog made me realize just how long its been since I’ve posted anything. This is just me saying that I’m sorry! I’ve been busy (duh) but didn’t really realize how much of my old development work was being sidelined until I finally started to catch up on my dev e-mails (4236 unread and counting….) So if I haven’t responded to something in the past month, I promise, I’m getting there. In addition, I just realized how horribly outdated local revisions/patches/branches become when you leave them unmerged/unattended for a month, so I have some real fun coming trying to get some of my cool metadata stuff in Beagle/Dashboard working against the current trunk.

Hopefully I’ll be back in a day or two with an awesome list of everything I accomplished/caught up on. However, its entirely possible that all I will be able to say is that I read and responded to my e-mail =/.

I know that the Beryl Project is fading from existence, but I still couldn’t help myself when I saw this enormous chunk of it at the Smithsonian.

Finally! Perhaps one of the last major features that Beagle needs to match Spotlight and Windows Live Search. Integration into every file choice made in the Gnome environment has long been a dream of mine, and a little less than a year ago, this bug was filed with hopes of making that dream a reality. After lots of hard work, a rough cut of the patch has been committed to the gtk+ trunk, with plans to add missing features in time.

While it will still be some time before this code reaches a Gtk+ release (development releases for the next major series start in late April) its great to feel some awesome love from our friends over at Gtk, a special thanks to Federico Mena Quintero for really being the driving force that finally got this done. What great is that plans for using Beagle’s index to help GtkFileChooser already seem underway.
For any budding Gtk+/Gnome/Beagle hackers out their, this feature still needs a lot of love. A list of bugs/todos exists here, and if someone were to get the ball rolling on anything on that list, it would be a huge help!

And last, the obligatory (if unexciting) Screenshots:

Here were searching for a Mail Attachment in Evolution.

And here we search for a file to upload in Epiphany:

I decided to give the Yahoo Publishers Network Beta a try, I never really make much money from ads, but it will be nice to have something to compare Adsense to.

So feel free to play around with the new ads, and my new analytics system (sorry if its slow!).

Its been an awesome week, its done justice to the last 6 months of blood tears and sweat that were poured into the great release that is Feisty Fawn.

A few things I wanted to mention, but first off, I should say Hi! My Ubuntu Membership was just approved, and I wanted to give a shout out and say hello to all those I haven’t met yet. Second, I wanted to do a little introduction, but given the general excitement, I figured I would link to my wiki and launchpad page, where most of you can get an idea of what I do, and where I can help. I’m always looking for new ways to get involved in the community, and if your working on something that you think I could help with, drop me an e-mail or find me on the IRC, and theres a good chance I’ll be all over it ;). That said I look forward to all the great releases down the road.

Just a few minutes ago I figured I would help the Ubuntu servers as best I could by seeding all the i386 iso’s, at the moment, there are close to 4000 seeds. I couldn’t ask the Ubuntu community for anything more, I downloaded the whole thing in close to 2 minutes, and am seeding now. Seeing as the bittorrent community needed no help, I decided that I would host the most popular iso on my site to help as best I can. (I thought about trying a full mirror, but this late, I figured it would do more harm then good, if I’m wrong, let me know!)

Here are the links, I know they aren’t for the main mirror site, but if you read my blog or one of the planet’s I’m on, and are thinking of grabbing the iso, consider using this link. If it doesn’t work, it means I could only be so generous, and I was going way over the 2TB limit this month, but I don’t think I’m that popular.

http://kubasik.net/files/ubuntu-7.04-desktop-i386.iso

http://kubasik.net/files/ubuntu-7.04-server-i386.iso

Technorati Tags: ubuntu, feisty

Powered by ScribeFire.

Ok, since my recent post on Dashboard there has been scattered interest in helping revive/work on Dashboard, so I wanted to post a list of things Dashboard needs before we can really look at making it usable at all.

• libdashboard

• Or as its better known, a C library that generates valid, parseable clues. This is not only critical since most plugin authors aren’t going to want to spend the time validating that mono can deserialize all their XML. But because we can generate bindings for most every other language once we have them in C.

• This is a huge one, Dashboard is still pre-alpha, and mostly proof of concept. While the code base could be brought up to production level, it needs tons of cleanup. This is where an army of open source dev’s can help the most, compile and install dashboard from SVN, then just play with it, and every time it crashes, track down why, and try to make it sane. This is long, slow, tedious, and thankless work, but it is an absolute necessity if people want to start using dashboard, as even simple race cases will crash dashboard most of the time.

• You have to be a little more familiar with the Beagle/Dashboard code base to help out with this, but we need them for a huge spread of plugins, and almost everything beagle can generate.

Just a note, I posted this list on the Dashboard Wiki.

Technorati Tags: linux, gnome, dashboard, mono, beagle, search

Powered by ScribeFire.

As I find myself getting more involved with the Ubuntu community and specifically the packaging of Beagle, I decided that it might be about time to apply for official Ubuntu Membership. I always feel a little odd in these situations, much the same as when I applied for membership to the Gnome Foundation. It feels weird trying to itemize and quantify my contributions, and moreover, I’m not a very good bragger when it comes to computers.

While the application process isn’t my favorite, once I commit to it, I find that there are plenty of tools that simplify the process, my blog not being the least of these, as its a nice reminder of what I have worked on. The harder part is deciding when its appropriate to apply, most guidelines on such topics are pretty vague, and to compound things, the most visible members tend to be hacker rockstars that are so above and beyond the accepted norm for membership that intimidation is almost impossible to avoid. I don’t mean any of this as a critique or attack, just my personal feelings and observations as someone slowly getting more involved in the community.

But I digress, The purpose of this post was to ask for anyone who was willing to give a quick glance over my wiki page at Ubuntu (its more or less your resume near as I can tell) to do so, and if willing, share their feedback. The next Community Council meeting isn’t until April 17th, so theres plenty of time. I would appreciate any of the following:

• Feedback on if I am applying too early (ie. not enough Ubuntu experience)
• Feedback on wiki page (formatting flukes, missed information, broken links etc)
• Anything I might have worked with you on, but forgot to include
• General Comments on life and/or its meaning

So yeah, I appreciate any feedback I can get!

Technorati Tags: gnome, ubuntu, linux, membership, gnome foundation,

Powered by ScribeFire.

Alright, before I get to the cool stuff, first things first, with current metadata system thriving,  and at Beagle’s current speed, if someone really wants to start to stabilize up the Dashboard API a little more, and start to make this something less abstract, IM,Call, Mail (either one), or even just show up in Cleveland, and I’d  be more than willing to help get Dashboard moving again.

Anyways, now that I’ve had another 2 minute excursion into the realm of getting Dashboard and its Banshee plugin building/working again, I have a quick 10 second screencast of dashboard being awesome.

http://qub333.googlepages.com/dashboard-banshee-plugin.ogg

The other cool technology that I really want to get integrated with Dashboard (or even Beagle) is the Open Natural Language Parser. In all honesty, there’s no  way to describe how awesome it is until you see it in action.

http://qub333.googlepages.com/opennlp-sharp.ogg

Sorry I’m too lazy to convert and all that jazz,  this is an impulse post that’s preventing work from getting done, so if someone wants to convert them, just let me know, I’ll host if you need.

So there’s been a lot of talk about a cool concept program coming out of the blur of Mugshot. While the exact goals of the project do seems a little more Web 2.0, the basic UI layout and goals seems to be remote services for something like Dashboard. Dashboard was the precursor to Beagle, (initially designed as an indexing and storage backed for Dashboard) and is basically a meta-clue processing center which brings up information relevant to whatever your are doing. The initial screenshots are extraordinarily old, but last summer during SoC, a complete redesign and recode took place, which is what can currently be found in the Google Code repo. While Dashboard is missing a lot of polish, I would recommend taking a look at its clue processing system, as its really quire ingenious, and the dream of every Beagle developer to eventually have to time to make it a real working solution again.

Anyways, I know its not the exact same, but its probably worth at lest a few minutes of playing time.

Sorry for the downtime this afternoon, had some php problems, luckily, sorting them out seemed to fix some other underlying issue, and the site is far more responsive. I hope no planets or readers got spammed in my attempts to fix everything!

So the past few weeks have seen me getting more and more frustrated with the performance lockups in Feisty. As a developer who’s been running it for several months now, I can say I’m very tolerant and never upset, just try my best to help, mostly by just finding my own solution. So cruft and buildup is inevitable, lots of random libraries compiled out of the packaging system to get something working last minute, and I know my kernel is a wreck, as its seen more than a few major upgrades. However, when I started to read benchmarks where Feisty was taking a significant beating in startup times, I started to get concerned. I agree with the reasoning that dropping the heavily optimized kernels was a good idea, I had stopped using them a bit before, due to my own stability problems.

Anyways, this prompted me to make good use of my 500 GB external drive, and I am backing up everything as I speak, then doing an all-out optimization of my system with no regards to consequence (just for some numbers/fun ) and then installing the beta fresh (on a LVM’ed hard drive). I hope to have something of a comprehensive optimization guide for Ubuntu after this, but whats most likely going to happen is I get ansey for the new Feisty and LVM

However, as I started my quest for a meaner and leaner Gnome/Ubuntu harmony, I trudged through the debates on prelinks, preloads, and readaheads (do what you will, nothing fantastic either way for me). But the first thing that truly changed my day to day desktop experience and in an incredibly positive way was Swiftfox. Its nothing revolutionary in terms of ideas, its just done well. Swiftfox really isn’t even a separate distribution of Firefox, its almost the exact same code, just with highly optimized releases for each processor. My initial thoughts were that it was going to be worthless, and maybe for your processor, it is, but with the Core Duo, its a world of difference. At first I assumed it was placebo, or something else I had done, but when my benchmarks showed no real differences anywhere else, Swiftfox outpaced Firefox, by a lot.

Just for fun, I have a few links below for rough web rendering performance benchmarks, make of them what you will, but let me say, drop your composting manager first, both beryl and compiz hated me when I tried, its too many rapid refreshes.

http://scragz.com/tech/mozilla/test-rendering-time

http://celtickane.com/projects/jsspeed.php

http://www.24fun.com/downloadcenter/benchjs/benchjs.html

Feel free to share your times, either way, I’m interested to know if this was a ‘My Firefox was so messed up’ or if Swiftfox is truly that awesome nerd inside that drives the speed of software.

Technorati Tags: mozilla, firefox, swiftfox, ubuntu, gnome, linux, optimize, speed

Powered by ScribeFire.

A follow up to my earlier post.

In my post last week I complained about how linux had no real torrent client that could go toe to toe with utorrent. Over the last week or so, I have been trying to follow up on all the awesome comments and suggestions that I got from readers. I found an answer.

Deluge is a Bittorrent client written in python and using libtorrent as a backend. Now, the release in Ubuntu Universe, and even its most recent 0.5 release were nothing to write home about, but the current SVN of Deluge shows serious promise. While I do *NOT* encourage users to try the SVN, as it is under active development (and I don’t want to force the project maintainer to answer a million questions about rapidly changing code) , I do feel that Deluge’s next release could match utorrent, or even surpass it! I found that while Deluge was taking a little longer to bootstrap less populated torrents, and the plugin/preference selection leaves something to be desired, Deluge is making huge progress in the speed area. Once I got torrents rolling, I found that Deluge was averaging about the same up and down as utorrent had been, and without the memory overhead of wine!

In short, I thank everyone for their awesome feedback, and thanks Deluge for stepping up and filling the niche for a full-featured and fast native linux client!

On a somewhat unrelated note, please consider filling out my survey on OpenID use if you haven’t already!

Miguel de Icaza has asked for some simple link help in getting Lame Blog some better Google results. I dunno if this helps at all, I don’t have a great Page Rank (I think its a 4 at the moment) so maybe its hurting, but the planets I get syndicated to do way better, so hopefully it helps.

Hey, so the 10 of you that actually visit my blog and don’t read it through a planet, you may have noticed me having some fun today. I decided that my blog was going to support OpenID, despite my hosting provider not having the needed math libraries, so my apologies if logins aren’t as speedy as you would like. However, dependencies and the like were not the hard part of the process at all, the OpenID Wordpress Plugin doesn’t really integrate into the site unless your running Kubrik or a derivative.

Luckily, there’s an easy enough solution, the OpenID API is pretty easy, and it didn’t take much to get everything (including Comments, and logging in via that link!) more or less running smoothly. On that note, don’t hesitate to nag, I hope its something people actually want to use. Its in that spirit that I have this little poll, no obligation, I was just thinking about pushing for Gnome to consider the MoinMoin plugin on live.gnome.org.

Ajax Polls and Planets do not like each other! Please participate in the poll here.

My last fun tidbit for the evening is actually really useful, if anyone runs a Wordpress blog, they know that its not the fastest or lightest thing in the world. But, most likely they have heard of wp-cache2, and have reached a comfortable compromise. One thing that really sucks about wp-cache2, is you can’t compress pages and use it, which means were missing out on the biggest web speedup there is. I realized this evening that you can totally utilize php’s native compression, and completely bypass the issue. If you add the following to your .htacces in the wordpress folder, your looking at a significant speedup, at least, I noticed one.

php_flag zlib.output_compression on
php_value zlib.output_compression_level 2

I believe the compression levels are 1-5, but don’t quote me, I’m using 2 for the moment, and I’m happy, I might do some more tweaking later though.