Elektrubadur

Converting Sources of This Blog to Org format

2023-05-28T00:00:00+02:00

It looks a bit bad that the last entry here was three years ago and was about how the site generation works, and here is another one on the same topic. However, it’s in the interest of making it more enjoyable for me to write more that I’ve switched from Zola to Hugo in order to be able to keep the sources in Org format.

Here I’ll just share the script I made to do the format conversion, in case it’s useful for anyone else. Actually, half the reason I’m even writing this entry is that I will need to later on repeat this process for another page I manage, and this way I have the script and some notes for then.

At the same time I’ve moved the sources from GitLab to Sourcehut [Update 2024-11-05: I have since moved them to GitHub].

I won’t delve deeper into the reasons for that, but I thought I’d post the Python script I used for the conversion in case it’s useful for anyone else. It’s not particularly pretty and has special casing for various things that were particular to my Zola theme, so if anyone else were to use it, they would in all likelyhood need to do some changes. I made a few manual fixes in the Org source after the conversion, of things that occurred to rarely to mke it worth automating.

The script does the following:

Clears out the target content directory. That’s the content/ directory under your current working directory.
Copy any non-Markdown files into the same location from the old content directory, assumed to be content.old/ in your current working directory.
Extract TOML front matter from the old source files and convert into Org front matter.
Run the remainder of the Markdown file through Pandoc to convert it from Markdown to Org.
Apply various substitutions to the body to fix some things Pandoc doesn’t handle, or just according to my preferences.
Write some metadata that can not be nicely encoded as Org front matter and put it in a config.toml file next to the Org file instead.

Then I just needed to go ahead and write a Hugo templates that works with this new content directory. Getting the same general structure of the page as in Zola isn’t difficult, as Zola is inspired by Hugo and uses a lot of the same conventions. I didn’t have to do much to be able to preserve mostly the same hierarchy of the source files for equivalent output.

If you are curious about the Hugo templates used to render this page, they are in the repository with the sources for this page.

Finally, if you want to use the script I’ve been talking about, you can download it here: convert.py

How This Blog Works

2020-05-03T00:00:00+02:00

This blog is created using the Zola static site engine. I use a combination of GitLab CI and cron to automatically upload it to my web server whenever I merge a change to the Git master branch of the home page.

The elektrubadur.se source code is kept here in GitLab [Update 2024-11-05: elektrubadur.se source code since moved to GitHub]. While writing a post I just run zola serve on my computer and can look at a dynamically updated website on a local web server.

Once ready, I commit on a branch and push it to GitLab, and then make a merge request. This might seem a bit silly for a one-man project, but this way I can disable pushes to the master branch entirely, which can help prevent an accident.

The repository includes the separate repository web-build as a submodule. It contains some scripts used to build and publish the web site.

My website is hosted at HCoop, a hosting cooperative. Because it uses AFS, passwordless login requires the use of Kerberos, so I can’t just push the page using a deploy key.

So, rather than pushing the website there, I’m pulling. I’ve put the script fetch_website.py from web-build on the target host, and makes it run it as a cron job:

*/5 * * * * k5start -qtUf "/etc/keytabs/user.daemon/$USER" -- "$HOME/scripts/fetch_website.py" "https://gitlab.com/bkhl/elektrubadur.se/-/jobs/artifacts/master/download?job=build" "$HOME/www/elektrubadur.se"

The k5start thing is another thing needed because of the use of AFS and Kerberos. The point here is that you give the script a URL to a GitLab artifact, and a target directory, and it will deploy it if there’s been any change.

The entire script is a bit long to post here, but it basically checks if there is a new artifact in GitLab for the website, built from the master branch, and if there is downloads and deploys it. It saves the ETag in a file, so that it can compare to that and skip downloading unless there has been a change.

Credits for tools used making this page, inspirations for the design &c. are also on About This Page.

Vimwiki + Nextcloud Notes

2020-05-02T00:00:00+02:00

I’ve been using Vimwiki for a while, which lets me keep interlinked plain text notes on my computer with minimal hassle.

For a while I’ve also kept these synced to my phone via the Nextcloud Notes application. I have a self-hosted Nextcloud instance, and simply point Vimwiki on my computer to the same directory in the directory that I have synchronized with Nextcloud.

The Nextcloud Notes app uses Markdown, so I have to configure Vimwiki to use that syntax, which is fine with me. One quirk of the Nextcloud Notes app is that rather than letting you name files independently, it will change filenames to whatever the first line of your note is. Until today that meant I had to keep making sure in Vimwiki to add those headers, but I’ve now finally automated that.

To start with, I added the function SetWikiHeader to my ~/.vim/plugin/config/vim.wiki:

if &runtimepath !~? "vimwiki"
  finish
endif

" My main wiki.
let g:vimwiki_list = [{
            \"path": "~/Documents/Notes/",
            \"syntax": "markdown", "ext": ".md"
            \}]

" Disable creation of temporary wikis.
let g:vimwiki_global_ext = 0

" Function to set header of current file to match filename.
"
" This helps interoparability with Nextcloud Notes, which assumes that
" the first line of the file matches the filename.
function SetWikiHeader()
    let first_line = getline(1)
    let filename = expand("%:t:r")

    if "# " . filename == first_line
        " First line is already correct.
        :
    elseif first_line =~ '^#*\s*\V' . escape(filename, '\') . '\m\s*$'
        " First line matches filename but has wrong header format.
        " Reformat to have proper format.
        call setline(1, "# " . filename)
    else
        " First line not matching filename at all. Prepend lines with
        " filename as a header.
        if first_line != ""
            call append(0, "")
        endif
        call append(0, "# " . filename)
    endif
endfunction

Vim script is pretty weird, had to search in documentation for how to do most things in there. Props to people who write entire programs in it, like let’s say those people that wrote Vimwiki.

Now, to make this run whenever I open a Vimwiki file, or before saving it, I put this in my ~/.vim/ftplugin/vimwiki.vim:

" Set header to match filename when opening or saving file, for
" interoperability with Nextcloud Notes.
call SetWikiHeader()
autocmd BufWritePre <buffer> call SetWikiHeader()

The end result is pretty nice. The only big catch is that I can’t rename files in Nextcloud Notes if they have a lot of links to them, since those links only get updated if I use the rename function in Vimwiki, but then that’s not something I do very often.

First Month with Fedora Silverblue

2020-02-08T00:00:00+01:00

I decided during my midwinter vacation that I’d try out Fedora Silverblue on my desktop, and I liked it, so here I am still using it on my work laptop.

Silverblue is a version of Fedora using rpm-ostree to provide a minimal and immutable base OS, providing a platform for running applications in containers.

Previously I was using Ubuntu with a lot of Ansible playbooks for installing things both globally and in my home directory. I used to keep those playbooks in the same directory as my dotfiles, but after the switch to Fedora I removed them. If anyone is curious about what that looked like, they are here.

The benefits I was hoping for was

Make my laptop setup more “out of the box”, to get away from having a big collection of Ansible roles for varied purposes, that all have to work together on my laptop.
Simplifying my base OS by moving development stuff and applications into containers.
Plain Gnome desktop without custom themes.
RPM-based. This is not an advantage in itself, but makes things a bit easier as I mostly deal with RHEL and Centos for work.

Applications

For applications, I’m using mostly Flatpak. I found after a while that applications I had installed from the Fedora Flatpak repository was more buggy than the ones I had from Flathub, so at one point I made a switch to get all my applications from Flathub.

Here’s the entire command line to get rpm-ostree into the current state of my laptop:

rpm-ostree reset

rpm-ostree install \
  fedora-workstation-repositories \
  ffmpeg-libs \
  google-chrome-stable \
  gparted \
  krb5-workstation \
  langpacks-en \
  langpacks-en_GB \
  langpacks-sv \
  langpacks-th \
  libgnome-keyring \
  NetworkManager-fortisslvpn-gnome \
  nextcloud-client \
  qemu \
  thai-arundina-sans-fonts \
  thai-arundina-sans-mono-fonts \
  thai-arundina-serif-fonts \
  vim-enhanced \
  vim-X11 \
  virt-manager \
  wl-clipboard

rpm-ostree install 'https://download1.rpmfusion.org/free/fedora/rpmfusion-free-release-31.noarch.rpm'
rpm-ostree install 'https://download1.rpmfusion.org/nonfree/fedora/rpmfusion-nonfree-release-31.noarch.rpm'

rpm-ostree install fuse-exfat exfat-utils

As you can see, there’s just a few things that has to be added as RPMs. Not a lot, considering that in the default installation, Silverblue just comes with a terminal, a file manager, and Firefox, in terms of applications.

Development environments

The main thing I’ve done so far to make it convenient for me to use as a development system is to create container images for use with Toolbox. I keep these images here in GitLab. I’ll write more about that in a future article.

For now I’ll summarize it by saying I’ve created some Toolbox environments for various languages, configured to keep caches and such in the container, rather than my $HOME, and it’s working quite nicely so far.

Gnome

I used the default Ubuntu desktop before, so this is not a dramatic change, but I like the look and feel of the default Gnome desktop. I also get a somewhat newer version of it, which is nice.

A few weeks after the switch, Tobias Bernard (Gnome developer) posted the article Doing Things That Scale, which resonated with me. While I’ve done some hacking to set up my development environments “just right” as described above, I’m trying to think of how to make that work more usable for others, maybe by contributing back to the Toolbox project, or by making some sort of GUI wrapper to manage a set of development environment toolboxes.

Simple Nextcloud To-Do List Automation with Python and WebDAV

2019-05-05T00:00:00+02:00

I keep a task list in a todo.txt format in Nextcloud. They are simply kept as two text files, todo.txt and done.txt in the root of my Nextcloud user directory.

Some clients already have the option to move tasks from todo.txt to done.txt when you mark a task as done. However, I would prefer they stick around in the todo.txt for a while, in case I change my mind.

At first I thought I’d just write a script to run as a cronjob on the server hosting Nextcloud, but that won’t work with syncrhonization, as it won’t let Nextcloud know the files have changed. I then thought there must be an API to read/write files to Nextcloud, and then quickly realized that this API is WebDAV. So, I should be able to make a Python script to do this, using requests. A more full-featured WebDAV library would be overkill for this.

Outline

Here’s what the script will do.

Get the todo.txt file.
If there are any completed tasks in there, get the done.txt file.
Strip the completed tasks from todo.txt.
Append the completed tasks to the content from done.txt.
Update todo.txt.
If there was no problem updating todo.txt, update done.txt. This step has to be retried in case of a conflict, because now the only copy of those done tasks is in the script’s memory.

This still leaves room for some issues if the script crashes in the middle of running, but I figured if it happens I’d get a mail from cron, and there’s a file history in Nextcloud so I can sort it out then. After running this a few months now it still hasn’t happened.

Implementation

Starting with imports and stuff of course:

#!/usr/bin/env python3

import argparse
import configparser
import re
import sys

import requests

In todo.txt, tasks are marked as done by prepending the token x at the start of the line. For this purpose, that’s really all we need to do, so I just create regexp to match these lines:

DONE_LINE_RE = re.compile(r"^x\b")

That’s it for the parsing.

Now, as this script will update two files, and I don’t want to lose any tasks, I need to worry about race conditions. In other words, I’ll need to use ETags.

For the structure of the script, I at first wrote it using the Plumbum CLI application, but later found that it was adding so little benefit I didn’t want that dependency. Instead I’ll use configparser and argparse for the same thing. For error handling, I’ll have functions return one of these values, and translate them to return codes at the end.

SUCCESS = object()
FAILURE = object()

It’s time for the main function.

def main():
    # Parse command line arguments
    argument_parser = argparse.ArgumentParser()
    argument_parser.add_argument(
        "-c", "--config", help="configuration filename"
    )
    arguments = argument_parser.parse_args()
    config_filename = arguments.config

    # Parse configuration
    config_parser = configparser.ConfigParser()
    config_parser.read(config_filename)
    try:
        username = config_parser.get("Connection", "username")
        password = config_parser.get("Connection", "password")
        hostname = config_parser.get("Connection", "hostname")
    except configparser.Error:
        sys.stderr.write(
            'Missing configuration in "{}".\n'.format(
                config_filename
            )
        )
        return FAILURE

    # Perform maintenance
    return archive_done_tasks(hostname, username, password)

This allows you to pass a command line argument pointing you to a configfile containing something like this:

[Connection]
hostname = your.nextcloud.instance.net
username = your_user_name
password = v3ris3curep4ssworð

Now, the function that’s the meat of the script, which actually gets and updates the files, and moves the completed tasks over.

The first loop will get the todo.txt, then put it back after filtering out the completed tasks.

If the initial GET fails, it will return successfully, since that probably means the server is unavailable, and we are better off just waiting for the next time cron triggers the script.

The next loop will retry up to ten times, because at that point todo.txt has been updated, so we should try a bit harder to get done.txt into a correct state.

This function is a bit long and would benefit from refactoring. (Pull requests accepted.)

def archive_done_tasks(hostname, username, password):
    base_url = "https://{}/remote.php/dav/files/{}/".format(
        hostname, username
    )

    session = requests.Session()
    session.auth = (username, password)

    todo_url = "{}/todo.txt".format(base_url)
    done_url = "{}/done.txt".format(base_url)

    todo_put = None
    while (todo_put is None) or not (
        200 <= todo_put.status_code < 300
    ):
        todo_get = session.get(todo_url)

        # If the first GET fails, just return and wait for next
        # time script is run.
        if not (200 <= todo_get.status_code < 300):
            return SUCCESS

        todo_lines = todo_get.content.decode(
            encoding="utf8"
        ).splitlines(True)

        updated_todo_lines = []
        added_done_lines = []

        for line in todo_lines:
            if DONE_LINE_RE.match(line):
                added_done_lines.append(line)
            else:
                updated_todo_lines.append(line)

        if not added_done_lines:
            # Nothing to do.
            return SUCCESS

        todo_put = session.put(
            todo_url,
            "".join(updated_todo_lines).encode("UTF-8"),
            headers={"If-Match": get_etag(todo_get)},
        )

    done_put = None
    i = 0
    while (done_put is None) or not (
        200 <= done_put.status_code < 300
    ):
        done_get = session.get(done_url)
        if not (200 <= done_get.status_code < 300):
            continue

        done_lines = done_get.content.decode("UTF-8").splitlines(
            True
        )

        done_put = session.put(
            done_url,
            "".join(done_lines + added_done_lines).encode(
                "UTF-8"
            ),
            headers={"If-Match": get_etag(done_get)},
        )

        # Retry this a few times since we updated todo.txt
        #  and want to get done.txt into a correct state.
        i += 1
        if i >= 10:
            return SUCCESS

Here’s the helper function to get the etag from a requests.Response object. I found that the ETags from Nextcloud has a content type suffix, so had to make another regexp to remove that :

ETAG_RE = re.compile(r'^"(.*)-gzip"$')

def get_etag(response):
    tag = ETAG_RE.match(response.headers["etag"]).group(1)
    return '"{}"'.format(tag)

Finally, we call the main function like this. The SUCCESS and FAILURE objects are converted into status codes and the script exits accordingly.

if __name__ == "__main__":
    result = main()

    if result is SUCCESS:
        sys.exit(0)
    elif result is FAILURE:
        sys.exit(1)

Here is the entire script.

This is just a start, if you want more advanced automation, handy things that could be added include (again, pull requests accepted):

Parse tags indicating repeating tasks. When those are completed, move the current one to done.txt, but create a new task for the next occurance at the same time.
Change it to keep completed tasks in done.txt for a certain numer of days after the completion date.

Suphan Buri

2018-12-10T00:00:00+01:00

I went on a work outing to Suphan Buri a little while back, here comes some photos. Most of the pictures are from the buffalo themed resort where we stayed.

Buffalo

Deer

Accomodations

Peacock

River flowing past nearby Sam Chuk market

Traffic

2017-09-15T00:00:00+02:00

Lower Parel, Mumbai.

Gecko on Window

2017-05-31T00:00:00+02:00

Gecko climbing outside of my window.

Frog

2017-05-11T18:00:00+07:00

Bathroom decoration on Koh Lanta.

Cow

2017-05-11T12:00:00+07:00

Resort cow.

Ants

2017-05-08T00:00:00+02:00

Fruit on the porch attracting ants.

Hermit Crab

2017-05-06T00:00:00+02:00

Hermit crab on the beach.

Bat Yam at Night

2017-04-29T00:00:00+02:00

Bat Yam night time skyline.

Jaffa

2017-04-22T00:00:00+02:00

Port of Jaffa.

Pigeon on Pillar

2017-04-11T14:00:00+02:00

Pidgeon on stoa pillar.

Minaret

2017-04-11T12:00:00+02:00

Mosque of Omar minaret, Jerusalem.

Pigeon by Sea

2017-04-10T00:00:00+02:00

Pidgeon by the Eastern Mediterranean Sea.

Bat Man

2017-04-08T18:00:00+02:00

Bat Man in Bat Yam.

Paramotoring

2017-04-08T10:00:00+02:00

Powered paraglider over Bat Yam beach.

Bat Yam Beach

2017-04-07T00:00:00+02:00

Bat Yam beach.

ACME

2017-03-22T00:00:00+01:00

Selfie outside of ACME building, Bangkok.

Grasshopper

2017-03-20T00:00:00+01:00

Urban grasshopper, Bangkok.

Morning Coffee

2017-03-19T00:00:00+01:00

Morning coffee at hotel ฟ้าใส, Buriram.

Road Trip

2017-03-18T00:00:00+01:00

Isan road trip.

Road Trip

Road Trip in Rain

Strawberries

2017-02-13T00:00:00+01:00

Brought some strawberries home from Korea.

Turtle

2017-02-11T00:00:00+01:00

Turtle at Sealife aquarium, Busan.

Cold Beach

2017-02-10T00:00:00+01:00

Busan beach in winter.

Spillt Coffee

2017-01-12T00:00:00+01:00

Spillt coffee on my desk.

Street Map

2016-12-19T00:00:00+01:00

My neighbourhood has these maps drawn on pavements, walls, street cabinets &c. I’m curious to know what they are for.

Rubber Band Ball

2016-10-27T00:00:00+02:00

Home made rubber band ball.

Water Fight

2016-08-18T15:00:00+02:00

Dogs playing in water.

Valley

2016-08-18T12:00:00+02:00

View of Ragunda valley.

Crows

2016-07-31T15:00:00+05:30

Crows by Mumbai promenade.

Cows

2016-07-31T12:00:00+05:30

Cows in Mumbai.

Bridge at Night

2016-07-11T00:00:00+02:00

Sydney Harbour Bridge at night.

Playground

2016-07-03T12:00:00+05:30

Playground in Mumbai.

Windy Beach

2016-07-03T09:00:00+05:30

Bandra beach, Mumbai.

Football

2016-06-22T00:00:00+02:00

Osotspa football club fans visiting Muangthong.

Gecko

2016-05-12T00:00:00+02:00

Gecko in paper envelope.

Swim

2016-05-02T00:00:00+02:00

With my parents on Koh Samet.

Greenery

2016-04-03T00:00:00+02:00

It is nice that my new flat has a bit of a view.

Muay Thai Night

2016-04-02T00:00:00+02:00

A photo I took while waiting for someone at the National Stadium station.

Fire Extinguisher

2016-01-29T00:00:00+01:00

Fire extinguisher on public van.

Loi Krathong

2015-11-26T00:00:00+01:00

Celebrating Loi Krathong.

Toast

2015-11-08T00:00:00+01:00

Toast with coin for scale.

Lake

2015-10-11T00:00:00+02:00

View from my parents’ cottage.

Fogged up

2015-10-05T00:00:00+02:00

Fogged up glasses.

China

2015-09-27T00:00:00+02:00

Collage of a trip to China.

Peafowl

2015-08-17T00:00:00+02:00

Peafowl on Koh Munnork.

Grill

2015-07-25T00:00:00+02:00

Grilling seafood.

Seafood

2015-06-30T00:00:00+02:00

Seafood, ready to be grilled.

Rain

2015-06-16T00:00:00+02:00

View from my office, with rain coming.

Elephants

2015-06-08T00:00:00+02:00

Painted elephants.

Dog on Back Seat

2015-05-16T00:00:00+02:00

Dog riding on motorcycle back seat.

Singapore

2015-04-05T00:00:00+02:00

Singapore skyline.

Racks

2015-04-04T00:00:00+02:00

Some server room.

Open Mall

2015-03-29T00:00:00+01:00

Open-air mall in Bangkok.

Van

2015-03-27T00:00:00+01:00

Inside of public van.

Sunbird

2015-03-26T00:00:00+01:00

Olive-backed sunbird nesting on my balcony.

Soekarno–Hatta

2015-03-12T00:00:00+01:00

Soekarno–Hatta International Airport, Jakarta.

Jakarta

2015-03-09T00:00:00+01:00

Jakarta skyline.

Corridor

2015-03-08T00:00:00+01:00

A corridor somewhere.

Semla

2015-03-01T00:00:00+01:00

Semla at Hotel Rembrandt, Bangkok.

Pole

2015-02-22T00:00:00+01:00

Painted pole by beach on Koh Samet.

Yard

2015-02-21T00:00:00+01:00

Yard of resort, Koh Samet.

Mall

2015-02-14T00:00:00+01:00

Inside of a mall.

Mall

Another Mall

Coffee

2015-01-04T00:00:00+01:00

Coffee in painted cup.

That Look

2011-01-18T00:00:00+01:00

This is an update on my ongoing project of close-up portraits made with my ЛОМО Т-43 40 mm f/4 lens, mentioned here. I have a fair collection of portraits by now, I few of which are good enough to probably make the final series. This might be my favourite so far. Watch this space.

Family Portrait

2010-12-06T00:00:00+01:00

A family portrait I made for a couple of friends.

Attention

2010-12-05T00:00:00+01:00

I have been working a lot in black and white portraits lately. Here is a vaguely Hitchcockian self-portrait from a couple of days ago.

Open

2010-12-01T00:00:00+01:00

A bit of honest street photography. I only had to wait a few minutes in front of this somewhat busy scene before the right moment appeared.

Horse 52375

2010-11-21T00:00:00+01:00

Anonymous horse.

Giant

2010-11-11T00:00:00+01:00

I photographed this river valley on my way home from Jämtland, a couple of months ago.

Agonizing

2010-11-05T00:00:00+01:00

I just acquired a ЛОМО Т-43 40 mm f/4 lens. It’s a very nice glass lens from the Leningradskoe Optiko Mechanichesckoe Objedinenie. Sadly, the farthest I can focus it on my SLR without colliding with the mirror is a couple of decimeters, giving this field of view. This self-portrait is the start of a project I plan to undertake with this very limited setup. More on that later.

Duet

2010-10-23T00:00:00+02:00

A playful portrait of a friend meant for advertising her music business.

Flautist Amputee

2010-10-17T00:00:00+02:00

A fountain sculpture in the Uppsala town garden.

Tied Together

2010-10-10T00:00:00+02:00

I went for a photographic dog walk with an acquaintance. Here posing on a bicycle path on Uppsala’s east side.

An Education

2010-09-17T00:00:00+02:00

Friend’s daughter reading an improving comic book.

Aerial

2010-09-16T00:00:00+02:00

Jackdaws flocking at dusk.

Rebecka, Rebecka, Rebecka

2010-09-14T00:00:00+02:00

Another composite with a small acquaintance.

Receiving All Things She Never Departs from Her Own Nature

2010-09-12T00:00:00+02:00

A composite group portrait, made in the botanical garden of Uppsala.

Off-road

2010-09-11T00:00:00+02:00

Me and a sporty French car on a field outside of Uppsala.

Lost and Found

2010-09-07T00:00:00+02:00

This is a line for hanging lost gloves and such in Uppsala’s English garden. While I was taking the picture a lady took time out of her day to stay and point out to me what an eyesore the arrangement is.

The Light of Learning

2010-09-03T00:00:00+02:00

Portrait with Uppsala’s grand university library, taken with a make-shift tilt-shift lens, made from a regular 50 mm lens and som bicycle inner-tube.

White Water and Colourful Mist

2010-08-30T00:00:00+02:00

Hällingsåfallet is a waterfall in northern Jämtland, falling into a canyon, probably formed by the end of the last ice age.

Unknown Soldier

2010-08-27T00:00:00+02:00

By a lake in Jämtland, while waiting for a fish to bite. Made with two stacked 50 mm lenses

Small Country Church

2010-08-25T00:00:00+02:00

The 15th century church of Ragunda parish, Jämtland, through an aperture in its bell tower.

Flowering Head

2010-08-18T00:00:00+02:00

Man and flower in the Uppsala botanical garden.

Flocking

2010-08-11T00:00:00+02:00

Jackdaws flocking over Uppsala.

Harvest

2010-08-07T00:00:00+02:00

The bumblebees favour these blue flowers in Uppsala’s Botanical Garden.

Sales Pitch

2010-08-04T00:00:00+02:00

My friend Zrajm wants to sell you his pocket dictionary of Klingon. Clearly a must have for anyone with too much time on their hands.

Sheep 641

2010-08-03T00:00:00+02:00

I did this long hand held capture using the technique of zooming in during the exposure.

Arch

2010-07-28T00:00:00+02:00

I did this long hand held capture using the technique of zooming in during the exposure.

In the Pink

2010-07-26T00:00:00+02:00

This was made looking for pink subjects for an assignment.