You are viewing msevior

Discovering Physics, Hacking AbiWord
 
[Most Recent Entries] [Calendar View] [Friends]

Below are 20 journal entries, after skipping by the 20 most recent ones recorded in msevior's LiveJournal:

    [ << Previous 20 -- Next 20 >> ]
    Monday, February 9th, 2009
    3:40 pm
    February 7th, 2009
    February 7th, 2009 was the day I turned 50. It was also the hottest day ever recorded in Melbourne, reaching 46.4 degrees, a good 0.8 degrees hotter than the previous record set way back in 1939. I was planning to writing a long blog post about this but it feels so wrong.

    On February 7th, Australia suffered its biggest natural disaster on record from unbelievable bushfires. Over 126 people are confirmed dead already. My sister lost her beloved home in the once-beautiful little town of Marysville. It is totally gone now.  

    Tuesday, October 7th, 2008
    11:30 am
    Thursday, September 11th, 2008
    11:20 am
    Turn On
    Last night my time, the Large Hadron Collider at CERN successfully circulated beam in both directions. We also recorded beam-gas interactions in the ATLAS detector. So all in all it was great start to a grand experiment which has been under development for 20 years.

    I woke this morning to find both Google and xkcd joined in the celebrations.





    Congratulations everyone!

    Wednesday, August 13th, 2008
    1:06 pm
    Stupid script to kill evolution
    Evolution has a number of great features that make it the best way to interface to our stupid MS Exchange server at work. However it does tend to hang on average once or twice a day while moving email back and forth to the Exchange IMAP service. At this point I find I need to find and kill all the separate processes used by evolution and restart. This got tedious enough that I wrote a stupid python script to do the work. 
    ------------------------------------------------------------------------------------

    #!/usr/bin/python
    #
    # Remove all instances of evolution
    #
    import os
    os.system('ps -aux | grep -i evo | cat > /tmp/killevo.txt')
    fs = open('/tmp/killevo.txt')
    ll = fs.readlines()
    for l in ll:
        bignore = 0
        l.expandtabs()
        lc = ''
        bsp = 0
        for c in l:
            if((bsp == 0) or (c != ' ')):
                lc = lc + c
            if c == ' ':
                bsp = 1
            else:
                bsp = 0
        ls = lc.split(' ')
        pid = ''
        for ss in ls:
            if ss == 'grep' or ss == 'emacs' or (ss.find('killevo') > -1):
                bignore = 1
        pid = ls[1]
        if bignore == 0:
            kills = 'kill -9 '+pid
            print 'Executing.. ',kills
            os.system(kills)
    fs.close()
    os.system('rm /tmp/killevo.txt')

    -------------------------------------------------------------

    I'm 100% certain that a real python/perl/bash hacker could this in 1/3rd the LOC but hey it works :-)
    Tuesday, July 29th, 2008
    3:11 pm
    Stealing ideas and http://abicollab.net in action
    Due to an overwhelming response from our users we've stolen reimplemented user colorization for abicollab. See the screenshot below. Colors for different users in the collaboration can be toggled on and off as needed.

    In addition I've made a little screencast showing http://abicollab.net in action. Note that collaborating users see changes and connect immediately just by openning the desired document. Click on the image to see it working.



    (Warning, huge, (78 MB) ogg file.)
    Monday, July 21st, 2008
    11:24 am
    James Denton GSoC progress
    I'm mentoring James Denton's Google Summer of Code project which is to enable multipage layout for sufficiently wide pages in AbiWord.

    Here is a nice screenshot showing the feature so far.




    James Denton's MultiPage July20_2008
    James Denton's MultiPage July20_2008

    James Denton's MultiPage Google Summer of Code project as of July 20, 2008

    Great stuff James!

    Saturday, July 19th, 2008
    11:38 am
    Netbooks on GNOME radar?
    I read with interest Chris Blizzards post and I scanned through the GUADEC talks. One topic which did not seem to be much talked about were netbooks like the EEE PC. This new market segment expects to have 10 million units this year with hundreds of millions over the next few years. A large fraction of these will run Linux.

    Is it just me or does this not seem a like HUGE opportunity for GNOME? We've concentrated on making a simple, easy to use desktop, why don't we have hackers jumping up and and down with excitement at the real opportunity to put our technology into the hands of hundreds of millions of people? Isn't this what we've wanted to do from the start?

    This is certainly a direction we're keen on taking AbiWord + http://abicollab.net.

    Is there a cabal somewhere that is plotting how to make use of this massive opportunity that I'm not part of?

    I must admit I'm mystified that at this moment when we finally have some genuine desktop traction, people are talking about decadence. Geeze lets make great apps and keep our desktop stable so ISV's will use us for their products.
    Thursday, July 17th, 2008
    2:00 pm
    Yet another opinion on GTK-3.0
    From the point of view of AbiWord, the big issues with GTK+ and GNOME are fonts, complex text, Input Methods, printing and precision positioning of text on screen.

    In principle all these are being addressed in the current GTK+2.x series, along with font-config, pango and cairo.

    It's not clear that GTK+ 3.0 will make fixing any of this easier. If it will please let me know. Otherwise I agree with Miguel, Morten and HP. Why make the development of large applications harder than it already is?

    BTW I have not given up on GNOME as a workstation and I especially feel that with small netbooks running linux appearing at the rate of tens of millions per year, that now is exactly the wrong time to do so.
    Wednesday, July 9th, 2008
    10:16 pm
    http://abicollab.net beta
    As Marc recently posted, our http://abicollab.net beta is now accepting people interested in in testing the abiword/web service we've created. This service allows users to easily setup and maintain a collection of collaborative created documents. Users can simulatanously edit these documents via single click on the website or by connecting directly through their local abiword instance. Perhaps the most practicle outcome of the service is that if you place a document on the webservice you can always be certain of seeing the most recent version of the document because connecting with abiword starts up a collaborative session. So there is no need to wait for your friends and colleagues to commit their changes to a document before you can see what they're doing. You see their changes in real time and can make your own changes as well.

    We're really excited about this and we believe we're created a unique service that provides users with a significantly improved user experience over Google docs and MicroSoft sharepoint. Which are our most obvious competitors.

    One additional feature Marc has just added is full, end-to-end encryption between abiword connections. With this the chances of someone being able to either snoop or inject false packets into your session is virtually zero.

    To get the full experinece you'll need the latest stable version of abiword from svn. Here are Marc's instructions for how to build it.

    Compile AbiWord 2.6.x, together with the AbiCollab plugin (make sure you have at least the libsoup, gnutls and asio developer packages installed;)

    svn co http://svn.abisource.com/abiword/branches/ABI-2-6-0-STABLE/ abiword-2.6
    svn co http://svn.abisource.com/abiword-plugins/branches/ABI-2-6-0-STABLE/ abiword-plugins-2.6
    cd abiword-2.6 && ./autogen.sh --prefix=<your_prefix> && make && make install
    cd ../abiword-plugins-2.6 && ./autogen.sh --prefix=<your_prefix> \
    --with-abiword=../abiword-2.6 --disable-all --enable-abicollab \
    --with-abicollab-service-backend && \

    make && make install

    Then you need to register at http://abicollab.net and for that you'll need a key.

    Our first batch of keys went very quickly. Here is another set for people who missed out yesterday.


    063EC8D1BC1C10F4-89A99EDA-DEAD9AC0
    <input ... > <input ... >
    7112E7F6F22296CF-2960B8D4-44FB26E9
    <input ... > <input ... >
    76CE17921EA66232-EC4E20AF-06ACF8C2
    <input ... > <input ... >
    994E6F6A7CD8E44A-3988E534-60F67BBB
    <input ... > <input ... >
    9B9A77A3E9EB1AC2-07738AF4-C2656C30
    <input ... > <input ... >
    CE10ADEB25805D44-64E370E9-6CC59FE8
    <input ... > <input ... >
    E9F36334BFC80F83-1FDA7032-0965E706
    <input ... > <input ... >
    EFE57E08871D6CEC-03FC8709-1D968A93
    <input ... > <input ... >
    F651E2852A789E2C-9506FD36-5D3E2051
    <input ... > <input ... >
    0393BC7C5B04E830-77438F03-261B1E01
    <input ... > <input ... >
    1B2A481A2259D9BE-6B1F65BC-DB9C6ED6
    <input ... > <input ... >
    4BEB4027ECB892B5-DB37B2FA-BB12CF33
    <input ... > <input ... >
    60229B94B8ED1C7A-672E48AD-64849FBF
    <input ... > <input ... >
    9CF4399815E9C0F1-C54AB9BF-86D0BF4A
    <input ... > <input ... >
    06B32EFFDBF61E02-070A20A2-D7531825
    <input ... > <input ... >
    12FFCA08CF7E2994-B737803F-8504074F
    <input ... > <input ... >
    23C57A0348E70F8B-6FF5DFC2-494A978B
    <input ... > <input ... >
    25509C7C546D8569-51A0D9F2-B4267FD1
    <input ... > <input ... >
    30BC52CA03689704-0FD6E3A0-D224D51A
    <input ... > <input ... >
    34CBDA2708AD2419-20F8D735-2FB912EA
    <input ... > <input ... >
    358E15920ED6BFCE-5D23FD49-61D9A49A
    <input ... > <input ... >
    3FB6015343494E17-31F1A470-998EC0EC
    <input ... > <input ... >
    4618864EA441F273-9D2068F6-C9C85CD6
    <input ... > <input ... >
    48846CF8C4A60A1B-B827535A-42912533
    <input ... > <input ... >
    607B56F6AE60E5AA-E06CBBAC-BEC0F2B9
    <input ... > <input ... >
    66B3C647A7340899-FCB3A952-1C347C56
    <input ... > <input ... >
    80B15C2DB0B511BA-A52BB32E-8AFFBF4B
    <input ... > <input ... >
    9886A8F476663CA1-881CD671-846E3AEE
    <input ... > <input ... >
    9CFAF4A89507D59C-9E0FCCFC-FEE67799
    <input ... > <input ... >
    A74D14E4904002AA-7D1ED424-85400D71
    <input ... > <input ... >
    AAF5314007CC27AD-C51AEBB5-271F800E
    <input ... > <input ... >
    ABC9419DE1C92656-0DDE0A07-433212E3
    <input ... > <input ... >
    B885AD1EE228BEC1-58A17074-879ED9A7
    <input ... > <input ... >
    FD0F9CDE7917A087-118A2E01-DD0AE144
    <input ... > <input ... >
    Thursday, June 26th, 2008
    7:58 pm
    Fedora 9 on the desktop
    I recently upgraded my desktop at work to an intel duo 3 Ghz, 3 GB RAM, 500 GB disk, Nvidia graphics and lovely 22 inch 1920x1200 resolution display. I promptly installed Fedora 9 with livna rpms. The whole thing works wonderfully well. Compiz-fusion plus gnome is a joy. In all the desktop experience is substantially higher than my previous 6-year-old system.

    I haven't used OSX but it's hard to imagine it getting better than this and I frankly don't understand why Free Software hackers want to use it. If you must use MS Office or the usual Windows apps, Wine works great! I have MS Office 2003 for the rare occasions when AbiWord and Gnumeric don't provide sufficient interoperability.

    My laptops DVD drive is misbehaving so I did a network upgrade to fedora 9 via preupgrade. This Just Worked. No mucking around with anything so far.

    All in all very impressive. Congrats to all the Free Software hackers who have given me this system! Keep up the good work and don't worry about decadence. This is what I and I'm sure most knowledge workers want. A system that just works, keeps out of your way and lets you do your job.
    Thursday, May 29th, 2008
    9:12 pm
    High Petrol prices
    There is a lot of noise in Australia now about high petrol prices. Neither large political party appears to have the courage to tell people what is really going on.

    If you would like to know why Petrol or Gasoline prices are very high and likely to keep increasing, read the following article by  Euan Mearns at this link.

    http://europe.theoildrum.com/node/4007
    Friday, May 9th, 2008
    10:25 pm
    Nice interview in redhatmagazine

    In which our heros (mostly uwog) talk about all things AbiWord.


    http://www.redhatmagazine.com/2008/05/08/abiword-team-interview/


    </shamlessplug>


    Thanks mether!
    Monday, May 5th, 2008
    11:25 pm
    AbiWord-2.6 packages for Ubuntu.
    Despite his heroic efforts, Ryan Pavlik has been unable to get AbiWord-2.6 into the official Ubuntu repositories.

    Nevertheless Ubuntu users can grab Ryan's packages for AbiWord-2.6 by following these instructions
    Tuesday, April 22nd, 2008
    12:46 pm
    Latex rules!
    To follow up on Marc's post,  Google Summer of Code students really wanted to work on our Latex exporter. I think we had 10 applications to enhance that feature but we had none to work on ODF.

    Another example that Words != Action in the Open Source world.
    Sunday, April 13th, 2008
    3:29 pm
    Pango In Action
    Thanks to Sankarshan Mukhopadhyay who recently provided a nice example of how far the freedesktop community has come in supporting complex text such as Indic scripts. I simply copied his post and pasted it into abiword and got the following.




    This was just copied from Sankarshan's blog post in firefox and pasted into abiword-2.6. Congrats to the pango hackers and Tomas Frydrych who made this "just work" :-) (Plus it's a nice example of the quality of our X11 clipboard support :-)

    Update:
    Changed the screen shot to show the font Lohit Bengali

    Saturday, April 12th, 2008
    7:17 pm
    All the cool kids are doing it...
    [msevior@seviorlap ~]$ history|awk '{a[$2]++ } END{for(i in a){print a[i] " " i}}'|sort -rn|head
    220 cd
    147 make
    137 gdb
    120 ls
    46 svn
    43 abiword
    35 loccp
    33 emacs
    31 ssh
    18 abib

    "loccp" essentially runs grep over *.cpp
    "abib" does cd /home/msevior/abidir/bin

    Yep I develop abiword in 15 gnome-terminals. Who needs a fancy IDE?
    Saturday, April 5th, 2008
    4:46 pm
    Video of my AbiCollab presentation
    Last Tuesday I gave my abicollab presentation at the Linux Users group of Victoria. The presentation was recorded and is now on the web in the place where my LCA 2008 presentation was meant to be.

    The audience was great and asked me lots of insightful questions. Unfortunately my inexperience at being recorded was obvious as I often forgot to repeat the question for the recording. I also spent a lot of time explaining our upcoming web service to support abicollab but unfortunately I couldn't demonstrate it. My laptop refused to work with the video projector. The first that that has happened to me in 6 years :-(

    A special thanks to Peter Lieverdink who organised things.

     
    10:59 am
    Ubuntu - Suffering Sclerosis
    Well I guess it is almost inevitable. After a few years as being first on ball and dynamic as anything, it appears Ubuntu is suffering the Sclerosis its supporters have accused other distros of having. Case in point. Look at the officious bureaucracy we're dealing with trying to get abiword-2.6 into the next Ubuntu builds.

    It is worth noting that Fedora, SuSE, Foresight Linux and Mandriva have all managed  to get AbiWord-2.6 into their distros. These all have similar release dates to Ubuntu.

    Fedora and other distros have developed ways to be far more dynamic in their support of upstream projects. Ubuntu should take a look to see if they can learn something from them.

    As it stands Ubuntu users look to be frozen out of the next wave of collaborative document creation available to Windows and other linux users. It's all pretty ironic since the initial impetus for abicollab came from a former Canonical employee, Jeff Waugh.
    Wednesday, March 26th, 2008
    5:37 pm
    AbiWord Back for GSoC in 2008
    If you're a student, interested in getting paid to work on the project that provides "Word processing for Everyone", then head on over to the Google Summer of Code homepage and write up an application for AbiWord.

    You don't have long! Applications close on March 31st!

                     
    Sunday, March 23rd, 2008
    2:45 pm
    A hint of Physics beyond the Standard Model of Particle Physics.
    Last Thursday, Nature published a result from my experiment, Belle that shows an unexpected imbalance in how matter as opposed to anti-matter behaves in nature. Our paper was covered by slashdot.

    The reason for the widespread interest is that it could be the first clue to both solving a puzzle that originated with the Big Bang Model of Universal creation and for a deeper level of understanding of particle physics than that provided by the standard model of particle physics.

    The standard model assumes that there are 3 generation of quarks (fractionally charged particles) and leptons (integrally charged particles).


    Source: http://www.benbest.com/science/standard.html

    Interactions amongst these particles are mediated by the strong, electroweak and higgs fields. Of the particles shown above, only the Higgs particle has not yet been discovered. Each quark and lepton also has a corresponding anti-particle with identical mass but opposite charge. The are many general descriptions of the standard model on the web but I'll summerize the relevant points. It works extremely well but it is highly unlikely to be the final word on how particles interact. There are almost certainly new principles in nature that operate at very small distance scales that are not incorporated in the standard model. There are 3 powerful pieces of evidence for this.

    Firstly: the masses of the neutrinos are much too small to be generated from the Higgs field.
    Secondly: calculations of the mass of the Higgs yield nonsensical results at energies above 1 TeV
    Thirdly: there is mass in the Universe.

    It is this last point that our result addresses. In the bIg Bang, particles and anti-particles were created in equal amounts. As the Universe expanded and cooled these annihilated one another and were transformed back to energy. In order for the matter that makes up the stars and galaxies to exist, there must have been an imbalance in how nature behaves with respect to matter and anti-matter. This imbalance would mean that a bit more matter survived this annihilation process and remained to form the stars, galaxies, planets and people of the universe. We call this imbalance "CP-violation".

    OK, now the standard model does provide a means for favouring matter over anti-matter, however it is too small by about 10  orders of magnitude. If the standard models was all there was, matter would not have been dense enough to form a single star.

    So here is one big hint for a new principle of physics. Whatever new principles of nature exist, one or more of them must provide additional CP violation. My experiment, Belle, has found many examples of CP violation in the decays of B-mesons. In the result cited in our paper we look at how the rates of:

    Anti_B0 => K -pi+ decays compared to B0 => K+pi- . We find that the former is 20% less likely than the latter. This clearly demonstrates CP violation happening in nature.

    In addition we measured the rates for:

    B- => K -pi0 and B+ => K +pi0 In this case the former process is about 14% more likely than the latter. The histograms showing these results are shown below and one can see by eye the imbalance in the numbers of observed events.



    source: http://www.nature.com/nature/journal/v452/n7185/full/nature06827.html

    The Standard Model can't predict absolute results for  these processes because of the  difficulties the theory has forming mesons from quarks. However the most obvious calculations do predict that the asymmetries should be the same. The fact that they're not is quite surprising. It is possible to stretch the hadronic corrections to the model so that they agree with the measurements but such efforts are both contrived and serve to make other measurements we've made disagree with calculations.

    So what does it all mean?

    A nice and far more detailed commentary on our results  is provided by  Michael Peskin in the same edition of Nature.

    http://www.nature.com/nature/journal/v452/n7185/full/452293a.html

    It could be that hadronic corrections will explain our result. However, the result could also be the tip of the iceberg. The first hint of some new principle in nature that was previously hidden from our view. If so, it has the profound effect of making life in this universe possible through ensuring matter is the predominant form of ordinary mass.
[ << Previous 20 -- Next 20 >> ]
My Website   About LiveJournal.com