It took me a while to find a used one for my size. Since I have quite long legs I wanted a frame height of at least 62cm. Unfortunately most of the bikes you find are more like 59cm/60cm. The one I finally found is 64cm so almost too big even for me. It's a black/white/pink Winora bike from ~1985 in very good condition, everything original. I'm really happy with it.

After a good clean with the help of YouTube (specifically GCN and Park Tools) I noticed that one of the Pedals is hard to turn. Since this was a "new bike" I decided to give servicing them a try. Again GCN and Park Tools to the rescue for removal and also for the servicing itself (GCN and Park Tools inkl. adjustent of Shimano SPD pedals).

Since I couldn't get the Pedals off at first here a few tips to get them off I gathered on various internet forums:

• Heat (Heat gun, etc. preferably applied to the crank)
• Ice spray (if available)
• Let it soak in WD-40 (or similar)
• Phosphoric Acid (Coke/Pepsi apparently, didn't dare to coat it in sugar water though)
• Longer/Special Pedal Wrench
• Hammer on normal wrench
• Percussion drill (to loosen the corrosion)
• Hammer on crank to loosen the corrosion
• Tenison belt between wrench and crank arm

Remember, that the left side (aka non drive train side) is the exact opposite of the "normal" direction. Always open the screws pushing the wrench towards the back of the bike. I was able to open mine after soaking it in WD-40 over night and some light-medium hits with a hammer on my normal 15mm wrench.

Don't forget to apply some Copperpaste/Anti-Seize or grease on the pedals before screwing them back in!

昨日、東京大学のZoomのオンライン授業を受けると、ドイツ人を見た。その後、宿題をする代わりにちょっと昼寝をしてマスクを買いに行った。薬屋に入るとイチゴが売られた。店を間違いましたね。

火曜日

Vｰ ます「～始める」、「～終わる」、「～続ける」。‌‌‌‌はっきりに言うときに使う。‌‌例：授業を参加し終わってから一緒に遊びませんか？

木曜日

V-る「～と～た。」、V-た「～たら～た。」。‌‌驚いたときに使う。‌‌例：窓を開けると雪が降っていた。‌‌

• What is the Device Tree in Linux?
• Useful Commands
• Writing your own Device Tree Overlays
• Example: Editing MCP2515 overlay to work with new SPI6

What is the Device Tree?

Roughly speaking the Device Tree describes all the hardware of a device in a tree format. It is especially used for anything that can't be discovered automatically like serial interfaces, internal memory, oscillators, ... (in contrast to e.g. USB or PCI devices). As the Raspberry Pi has a lot of GPIO headers exposed, additional devices can be added that the current device tree doesn't know about. That's where overlays come into play.

The default base device tree source file for the RPI4 B can be found in the kernel sources in the raspberrypi/linux under linux/arch/arm/boot/dts/bcm2711-rpi-4-b.dts. This is the source file for the board, which is based itself on the source file for the SoC linux/arch/arm/boot/dts/bcm2711.dtsi.

What is an overlay?

Overlays are applied to the base device tree at a later point to change it, usually to add a device, or to configure/enable one that is already present in the base device tree.

For example: the overlay spi5-1cs-overlay.dts configures the spi5 node already present in the device tree to enable it with one CS pin and a spidev device.

All available overlays can be found in the /boot/overlays directory. Albeit already compiled.

Useful Commands

dtc

dtc is the device tree compiler. It is used to losslessly compile .dts files to .dtb files. As a naming convention all overlays source files end in -overlay.dts and are compiled to *.dtbo.

pi@raspberrypi:~ $ dtc -@ -I dts -O dtb -o my_overlay.dtbo my_overlay-overlay.dts 

The -@ is needed for overlays as it prevents the compiler from looking up the references from the source file, which it - of course - wouldn't find.

dtoverlay

dtoverlay is used to dynamically load device trees.

• dtoverlay -a to list all available overlays
• dtoverlay -l to list the currently applied overlays
• dtoverlay -h <overlay_name> to get information about any overlay, its parameters and possible values
• dtoverlay <overlay_name> <param_name>=<param_value> to add an overlay with some parameters
• dtoverlay -r <overlay_name> to remove a previously loaded overlay

Note: dtoverlay -l only lists dynamically loaded modules and only those can be removed with dtoverlay -r. Anything that was loaded via /boot/config.txt is considered part of the base device tree. Some overlays have to be added this way and can't be loaded by dtoverlay.

raspi-gpio

• Install it if not yet present: apt-get install raspi-gpio
• raspi-gpio funcs lists all available functions for all pins
• raspi-gpio get lists how every pin is currently configured

Debugging

• Add dtdebug=1 to /boot/config.txt to get a detailed output from vcdbg log msg, when an overlay fails to be applied at startup it will just be skipped.
• Remember dmesg | grep -i <device_name> can be useful to see if the module corresponding to the overlay has been loaded successfully.
• dtc -I fs /proc/device-tree prints the current complete device tree

Writing your own Device Tree Overlays

For an explanation of the syntax of *.dts files take a look at the official documentation. It's really good and you'll need it as the syntax is weird. Especially for overlays.

Miscellaneous

The number behind the @ in the name of a node (e.g. can@0) is the address inside the parent the node can be found at. It is by convention equal to the value of the reg property, in this case reg = <0>;.

#address-cells = <1>; defines how many addresses the children contain (here 1).

#size-cells = <0>; sets the size of the address space to 0. That means that the children will have a single address, instead of an address space (for e.g. memory).

Combined one can conclude that the reg property will only contain one value for the address and an empty value for the length (like above).

This also means, that this overlay can only be enabled once on a given SPI bus, because the addresses would clash. To enable it a second time the reg property and address have to be changed. This is independent of the CS used.

Example: Editing MCP2515 overlay to work with new SPI5

The overlay mcp2515-can0-overlay.dts references <&spi0> in multiple places. In principle the only thing you have to do to make it work with another spi interface is to change all of them to e.g. <&spi6>. While you're at it also change all occurrences of can0 to can6 and can0_osc to can6_osc so they don't conflict when loading both.

/*
 * Device tree overlay for mcp251x/can0 on spi0.0
 */

/dts-v1/;
/plugin/;

/ {
    compatible = "brcm,bcm2835";
    /* disable spi-dev for spi0.0 */
    fragment@0 {
        target = <&spi0>;
        __overlay__ {
            status = "okay";
        };
    };

    fragment@1 {
	target = <&spidev0>;
	__overlay__ {
	    status = "disabled";
	};
    };

    /* the interrupt pin of the can-controller */
    fragment@2 {
        target = <&gpio>;
        __overlay__ {
            can0_pins: can0_pins {
                brcm,pins = <25>;
                brcm,function = <0>; /* input */
            };
        };
    };

    /* the clock/oscillator of the can-controller */
    fragment@3 {
        target-path = "/clocks";
        __overlay__ {
            /* external oscillator of mcp2515 on SPI0.0 */
            can0_osc: can0_osc {
                compatible = "fixed-clock";
                #clock-cells = <0>;
                clock-frequency  = <16000000>;
            };
        };
    };

    /* the spi config of the can-controller itself binding everything together */
    fragment@4 {
        target = <&spi0>;
        __overlay__ {
            /* needed to avoid dtc warning */
            #address-cells = <1>;
            #size-cells = <0>;
            can0: mcp2515@0 {
                reg = <0>;
                compatible = "microchip,mcp2515";
                pinctrl-names = "default";
                pinctrl-0 = <&can0_pins>;
                spi-max-frequency = <10000000>;
                interrupt-parent = <&gpio>;
                interrupts = <25 8>; /* IRQ_TYPE_LEVEL_LOW */
                clocks = <&can0_osc>;
            };
        };
    };
    __overrides__ {
        oscillator = <&can0_osc>,"clock-frequency:0";
        spimaxfrequency = <&can0>,"spi-max-frequency:0";
        interrupt = <&can0_pins>,"brcm,pins:0",<&can0>,"interrupts:0";
    };
};

I additionally edited the target=<&spi6> section and added a parameter so that the CS pin can be freely chosen as well:

/*
 * Device tree overlay for mcp251x/can6 on spi6.0
 */

/dts-v1/;
/plugin/;

/ {
    compatible = "brcm,bcm2835";
   
    /* spi6 cs pin config */
    fragment@0 {
        target = <&spi6_cs_pins>;
        frag0: __overlay__ {
            brcm,pins = <26>; /* gpio26, hardware pin 37 */
            brcm,function = <1>; /* output */
        };
    };

    /* spi6 config */
    fragment@1 {
        target = <&spi6>;
        frag1: __overlay__ {
            /* needed to avoid dtc warning */
            #adress-cells = <1>;
            #size-cells = <0>;

            status = "okay";
            pinctrl-names = "default";
            pinctrl-0 = <&spi6_pins &spi6_cs_pins>;
            cs-gpios = <&gpio 26 1>;
        };
    };

    /* the interrupt pin of the can-controller */
    fragment@2 {
        target = <&gpio>;
        __overlay__ {
            can6_pins: can6_pins {
                brcm,pins = <16>; /* gpio16, hardware pin 36 */
                brcm,function = <0>; /* input */
            };
        };
    };

    /* the clock/oscillator of the can-controller */
    fragment@3 {
        target-path = "/clocks";
        __overlay__ {
            /* external oscillator of mcp2515 on SPI6.0 */
            can6_osc: can6_osc {
                compatible = "fixed-clock";
                #clock-cells = <0>;
                clock-frequency  = <16000000>;
            };
        };
    };

    /* the spi config of the can-controller itself binding everything together */
    fragment@4 {
        target = <&spi6>;
        __overlay__ {
            /* needed to avoid dtc warning */
            #address-cells = <1>;
            #size-cells = <0>;

            can6: mcp2515@0 {
                reg = <0>;
                compatible = "microchip,mcp2515";
                pinctrl-names = "default";
                pinctrl-0 = <&can6_pins>;
                spi-max-frequency = <10000000>;
                interrupt-parent = <&gpio>;
                interrupts = <16 8>; /* IRQ_TYPE_LEVEL_LOW */
                clocks = <&can6_osc>;
            };
        };
    };

    __overrides__ {
        cs_pin = <&frag0>,"brcm,pins:0",
                 <&frag1>,"cs-gpios:4";
        oscillator = <&can6_osc>,"clock-frequency:0";
        spimaxfrequency = <&can6>,"spi-max-frequency:0";
        interrupt_pin = <&can6_pins>,"brcm,pins:0",
                        <&can6>,"interrupts:0";
    };
};

This just needs to be compiled and copied to /boot/overlays/:

dtc -@ -I dts -O dtb -o mcp2515-can6.dtbo mcp2515-can6-overlay.dts
cp mcp2515-can6.dtbo /boot/overlays/

After that you can load it dynamically:

overlay mcp2515-can6 cs_pin=16 interrupt_pin=26 oscillator=16000000

Or statically adding to /boot/config.txt. Don't add anything else and make sure that nothing uses the same pins!

dtoverlay=mcp2515-can6,cs_pin=16,interrupt_pin=26,oscillator=16000000

Now that the device is known to Linux it behaves the same as usual. Just follow any of the myriad of guides on connecting the MCP2515. Pay special attention the the mentioned common mistakes (like having only one node). Put the can interface up and start sending data:

ip link set can0 up type can bitrate 50000
cansend can0 111#FF

It could also be adapted to work for any SPI interface like the overlay for the MCP3008 works for SPI0, SPI1 and SPI2.

I hope this was somewhat useful.

More Resources

• Official Raspberry Pi Documentation on Device Trees
• Official Raspberry Pi Device Tree Forum
• Official DeviceTree.org specification
• Good Explanation on how to write your own Device Tree

In this article I will describe how I set up my mail on the wonderful Uberspace 7 using .qmail and maildrop to automatically create subfolders in my inbox and sort mail in these based on the recipient (i.e. my) mail address.

The basic idea is that if you own an domain like yourdomain.tld you can add unlimited email addresses. So I had the idea (which is not new) to create different email adresses for different services like facebook@yourdomain.tld for Facebook, github@yourdomain.tld for GitHub, etc. You get the idea.
Creating a new mailbox for every service is way too much work. Catching all emails in one mailbox and creating a new folder and a filter to sort the mails into a folder every time is still really annoying. It's way easier if you can just register for a site and a new folder in your inbox gets created automatically when an email arrives on the domain for the new service.

So first I created my main mailbox which I use for my own mail. Emails to this address go directly into the inbox without filtering.

uberspace mail user add isabell

Then I added yourdomain.tld to my uberspace for use with email. Don't forget to set the corresponding DNS records on your provider (will be output by the add script).

uberspace mail domain add yourdomain.tld

On Uberspace the first sorting is done by .qmail files. By default all emails that are not caught by some specific .qmail file will match .qmail-default where it gets send to |/usr/bin/vdeliver (The pipe at the beginning tells qmail to pipe the mail into an external program, in this case vdeliver). More on this in the Uberspace Wiki. vdeliver then delivers the mail to the corresponding inbox if it exists and drops all others (actually they are dropped earlier, but lets keep it simple)
Because we want to receive all emails we forward all mails that arrive at .qmail-default directly to our inbox

echo "isabell@yourdomain.tld" > .qmail-default

Now all emails arrive at your inbox. So where's the filter? Now we get to maildrop. Instead of accepting all mails directly into our inbox we call maildrop first. How? We tell qmail to do it for us:

echo "|maildrop $HOME/.myfilter" > .qmail-isabell

$HOME/.myfilter contains our filter configuration. For more information on maildrop filters see the Uberspace Wiki.
Take a look at the filter first:

logfile "$HOME/mailfilter.log"

# set default Maildir to user isabell
MAILDIR="$HOME/users/isabell"

DESTDIR="$MAILDIR"

# Instead of looking at "To" or "Cc" we look at the
# "Delivered-To: " header, which is added by qmail, because
# it's easier and it's reliably there in contrast to the others

# For now just loop over all occurences and take the last email
ADDR=""
foreach /^Delivered-To:\s*.*/
{
    ADDR=getaddr($MATCH)
}

# Use sed to extract just the local part of the url without username
SERVICE=`echo $ADDR | sed 's/^[a-z0-9]*-\(.*\)@.*/\1/'`

# If it was delivered to a subdomain, put it in a subfolder, otherwise
# deliver it to the main inbox (also in case of an error)
if ( $SERVICE )
{
    DESTDIR="$MAILDIR/.INBOX.$SERVICE"
}

# check if the folder already exists and create and subscribe if necessary
if ( ! `test -d "$DESTDIR"` )
{
    `maildirmake "$DESTDIR"`
    `echo "INBOX.$SERVICE" >> "$MAILDIR"/subscriptions`
}

# last line, deliver to $MAILDIR
to "$DESTDIR"

I already added a lot of comments, so I think most commands should be clear, but here again a basic rundown. qmail adds a Delivered-To: header when it delivers a mail to an inbox. It does not do this if you call a programm in the .qmail file (i.e. |maildrop .myfilter, I learned this the hard way). This is why we first catch all mails in .qmail-default and forward it to our main user .qmail-isabell and THEN call maildrop to apply the filter, instead of just calling maildrop directly in the .qmail-default.
Why go to such lenghts? Because the To: and Cc: headers are not always there (think mailing list, bcc, ...) and there can be multiple recipients. The solution via Delivered-To: is easier since there may be multiple headers, but the first one (i.e. first added, listed last) will be the one where the other server initially sent our email and it will be the only email in this header. Thanks to Thorsten Köster for the idea.

With that in mind the script is easy. First define a default mailbox and set the destination mailbox to to be sure that no emails are dropped. Next go over all Delivered-To: headers, get the last one and extract the local part (i.e. <address> below) with sed. Remember the domain looks like this:

Delivered-To: <user>-<namespace>-<address>@<domain>.<tld>

If you have a namespace enabled you probably need to adjust not only the filter, but probably also the forwarding earlier.

So if we have a valid subdomain (i.e. it wasn't sent to isabell@yourdomain.tld directly) we check if there the directory in the inbox already exists, if not create it and subscribe to it. Lastly deliver the mail.
Two things to note: Subfolders in Mailbox format are created with leading dots and dots as seperators and not a "real" directory tree. Second, in IMAP you have to subscribe to a folder on a client to be able to see it. Usually you subscribe to all folders when you add the account to your client, but the oftentimes don't get updated automatically if new ones appear, therefore hiding the newly created folder. But because the list is stored on the server (so it get's synced across clients) we can easily modify it.

That's it! Lot's of explanation and not so much to do, but great results.

Here's my guide on installing Isso on Uberspace 7 for comments als alternative to Disqus and Commento. See also my last posts on Installing Ghost and Installing Commento and my thoughts for choosing Isso over Commento on Comments on Ghost on Uberspace 7.

The official Install Guide of Isso suggests to use pip, in our case with the --user flag for a local install:

pip install --user isso

(If you like to support python3 you can also do a pip3 install --user isso or even use pip3.6)

As in the guide for Ghost we need to choose a port:

ISSOPORT=$(( $RANDOM % 4535 + 61000)); netstat -tulpen | grep $ISSOPORT && echo "versuch's nochmal"

We want notifications when someone comments on our blog. Therefore we create a mailuser to send the emails from and set a password (of course you can skip this if you don't need/want it):

uberspace mail user add comments

Now we create a config file in ~/etc/isso/. We chose to configure Isso with smtp mail, therefore we add the [smtp] section and notify = smtp to our configuration. You can just copy the whole cat command and paste it in the shell:

mkdir ~/etc/isso
cat <<__EOF__ >~/etc/isso/user.cfg
[general]
dbpath = $HOME/etc/isso/comments.db
host = https://yourdomain.tld/
log-file = /home/maxb/logs/isso_logfile
notify = smtp

[server]
listen = http://localhost:$ISSOPORT/

[smtp]
username = comments@$USER.uber.space
password = SECRETMAILBOXPASSWORD
to = yourmailadress@yourhoster.tld
from = comments@yourdomain.tld
host = $HOSTNAME
__EOF__

After executing don't forget to change yourdomain.tld to your actual domain and to enter the password you chose earlier (~/etc/isso/user.cfg). Https only is sufficient, because uberspace enforces https and redirects all unencrypted requests. Of course you omit the smtp and notify settings if you didn't create the mailuser. More on the server config in the Official Documentation.

To make isso available on the outside we need to add a subdomain. For example comments.yourdomain.tld:

uberspace web domain add comments.yourdomain.tld

Now we need to tell Apache to proxy all requests to Isso. We do this by creating a directory for the subdomain and placing a .htaccess file in there:

mkdir /var/www/virtual/$USER/comments.yourdomain.tld
cat <<__EOF__ >/var/www/virtual/$USER/comments.yourdomain.tld/.htaccess
DirectoryIndex disabled
RewriteEngine On
RewriteRule ^(.*) http://localhost:$ISSOPORT/\$1 [P,L]
__EOF__

The first row prevents Apache from serving an index.html when just accessing the domain. This would prevent the Isso client (i.e. your website) from reaching the right endpoint at the server.
The last line of the file redirects the traffic to the Isso server.

Lastly comes the service file. We need to create a service file again, like we did for Ghost:

cat <<__EOF__ >~/etc/services.d/isso.ini
[program:isso]
command = $HOME/.local/bin/isso -c $HOME/etc/isso/user.cfg run
autostart = true
autorestart = true
__EOF__

It's pretty self-explanatory. autostart starts the isso server on startup and autorestart restarts the server when it fails.

Now just reread the config and update it to start Isso!

supervisorctl reread
supervisorctl update

Lastly we need to add the comment section to the blog. Probably below every post, so we take a look at our theme, which is casper, and find the ~/ghost/content/themes/casper/post.hbs. Some way down is already an commented part for Disqus. There we can add (obviously not inside the commented part) the following snippet:

<script data-isso="https://comments.yourdomain.tld/"
    src="https://comments.yourdomain.tld/js/embed.min.js"></script>
<section id="isso-thread"></section>

Again, don't forget to change your domain!

That's almost it. Restart Ghost to see the comment section appear!

supervisorctl restart ghost

If you have questions, leave a comment below.

Yours,
Max

This post explains how to install Commento (current version 0.3.1) on a Uberspace 7 installation.

Prerequisites

First choose a random port that is still free

COMMENTOPORT=$(( $RANDOM % 4535 + 61000)); netstat -tulpen | grep $COMMENTOPORT && echo "versuch's nochmal"

Because there is no port forwarding as of yet I decided to use apache as a port 80 proxy to commento. So the first step is so register a new subdomain. I chose comments.yourdomain.tld. Register it with the uberspace tool and don't forget to set your DNS records:

uberspace web domain add comments.yourdomain.tld

Next create a folder in your virtualroot for your subdomain and

cd /var/www/virutal/$USER/
mkdir comments.yourdomain.tld
cd comments.yourdomain.tld

Place an .htaccess file in there with the following contents to proxy the requests to the commento server and don't forget to change COMMENTOPORT to the port echo $COMMENTOPORT gives you:

RewriteEngine On
RewriteRule ^(.*) http://localhost:COMMENTOPORT/$1 [P,L]

All of this will probably be obsolete in the near future.

Installing Commento

First create a subfolder for the binary and databasefile:

mkdir ~/commento
cd ~/commento

Next lookup the latest release url, download the latest release of Commento from Github and extract it.

wget "https://github.com/adtac/commento/releases/download/v0.3.1/commento_0.3.1_linux_amd64.tar.gz"
tar -xzf commento_0.3.1_linux_amd64.tar.gz
rm commento_0.3.1_linux_amd64.tar.gz

Now copy the assets (namely the javascript and stilesheet) into your ghost theme. For example I copy them in my casper theme:

cp assets/style/commento.css ~/ghost/content/themes/casper/assets/css/
cp assets/js/commento.js ~/ghost/content/themes/casper/assets/js/

Edit your post.hbs to include the following code snippet I wrote. In my case its the ~/ghost/content/themes/casper/post.hbs. In the casper theme there is already a commented out section for Disqus comments. Just place it right above/below that.

<h2>Comments</h2>
<div id="commento"></div>
<script
    src="{{asset "js/commento.js"}}"
    data-div="#commento">
</script>
<script>
    Commento.init({
        serverUrl: "//comments.yourdomain.tld",
        commentoCssUrl: "{{asset "css/commento.css"}}"
    })
</script>

Create the service file ~/etc/services.d/commento.ini with the following content and again don't forget to change it to your actual COMMENTOPORT:

[program:commento]
directory=/home/$USER/commento
command=/home/$USER/commento/commento
environment = COMMENTO_PORT="COMMENTOPORT"
autorestart=true

Reread, update and restart all services, so ghost also reloads the script:

supervisorctl reread
supervisorctl update
supervisorctl restart all

And you're done! There should be a comment section below every post now. If you have questions, leave a comment below.

This is just a small discussion on options to add comments to Ghost on your fabulous Uberspace. See also Ghost Install on Uberspace 7

tl;dr I wrote a Guide on Installing Commento on Uberspace 7 but then decided to go with Installing Isso on Uberspace 7 because it has support for notifications.

Options

First off: simplest is probably if you just use one of the commenting frameworks available which are explained pretty good on this site like:

• Disqus
• Facebook
• NodeBB
• Livefyre

If you don't want to rely on a seperate service, like your privacy (or don't like how slow they are) and are not particulary fond of hosting your own nodeBB there are other options:

• Isso Python + SQLite3 (3135 Stars)
• Commento Go + SQLite3 (2551 Stars)
• Schnack Node + SQLite3 (1196)
• juvia Ruby on Rails + PostgreSQL (1025 Stars)
• CoralProject Talk Node + MongoDB + Redis (872 Stars)
• Discourse Ruby on Rails + PostgreSQL + Redis (24666 Stars)

I included the Github stars (as of writing), to show that most of them are already quite established with Discourse being far ahead.

As juvia is currently unmaintained and PostgreSQL and MongoDB aren't currently available at Uberspace 7 I feel like the last three are all too much hassle. They're also kind of overkill for such a small and simple blog.

So let's take a closer look at Commento and Isso:

• Commento is faster
• Isso has one level of nesting only
• Both currenly use SQLite, but commento will implement others
• Isso is older (more mature?)
• Isso looks more like Disqus, which I like more.
• Commento has Akismet support
• Isso has voting functionality
• Isso has email notifications

Because I already found a Guide on Isso for Uberspace 7 I thought I'd try it with Commento.

Here is my guide for Installing Commento.

After getting it up and running though I decided that I wanted email notifications, which are currently not supported on Commento. Therefore I wrote a guide on Installing Isso.

Ich spiele weiterhin mit der Google Spracherkennung herum. Funktioniert immer noch sehr gut. Insbesondere gewöhnt man sich daran die Satzzeichen mit zu diktieren.

Im Dorm haben sie seit kurzem einen Plastik Bambusbaum aufgestellt in denen sie Glückwünsche und wünsche gehängt haben. Ich vermute einfach mal, dass es das Tanabata Festival ist, bin mir da aber nicht ganz sicher, da ich noch nicht dazu gekommen bin jemanden darüber auszufragen.

Weiterhin war ich heute mal wieder im Café und nachdem ich da meinen Eiskaffee getrunken hatte habe ich beim Rausgehen wie die anderen Gäste auch ein ブレゼント (Present) bekommen, das aus einem Rasierer, Shampoo und Waschdeo besteht. Fand ich sehr interessant..

Auch habe ich zum ersten Mal einen Reifen nicht abgeschlossen ohne Fahrrad gesehen, normalerweise ist es ja andersrum oder der Reifen ist angeschlossen. Gibt’s wohl nur in Japan.

Meine Schuhe die ich aus Deutschland mitgebracht habe, geben wohl auch langsam den Geist auf. Bei einem fehlt an der Ferse ein relativ großes Stück der Sohle. Keine Ahnung wie das passiert ist oder wo dieses Stück gelandet ist.

Ganz langsam aber sicher macht mein IMDb Datenanalyseprojekt Fortschritte. Wer Lust hat kann sich das ja mal auf Github anschauen; da lade ich ab und zu mal den aktuellen Stand hoch. Freue mich über Kommentare.

https://github.com/M4a1x/IMDBanalysis

Tja die Posts werden wohl langsam wieder spärlicher, das liegt allerdings aber auch daran dass in letzter Zeit nicht viel passiert ist. Heute habe ich nicht viel gemacht. Heute morgen etwas Mario gespielt auf meinem 3DS, danach ein wenig japanisch & programmieren und anschließend Wäsche gewaschen. Dann hat mich Erika angerufen und gefragt ob ich Zeit habe beim Bambus schneiden zu helfen das war so gegen 2 Uhr. Etwa gegen 6 Uhr war ich dann wieder zurück im Dorm. Die meiste Zeit hat es eh geregnet, von daher war auch nicht viel mit Ausgehen, Wandern oder ähnlichem. Heute Abend habe ich Spotlight geschaut. Verdammt guter Film. Und anschließend ist mir die Idee gekommen mal eine diktier App für’s Handy auszuprobieren. Dieser Text z.B. wurde mit Google voice input diktiert. Funktioniert insgesamt ziemlich gut ist nur irgendwie ungewohnt Punkte und Kommas zu diktieren. Wobei ich Kommas meistens vergesse. Bin immer noch da hin und hergerissen ob ich das Datensammeln von Google aktivieren soll um die Spracherkennung zu verbessern, oder ob sie eigentlich schon gut genug ist. Ich sollte unbedingt noch weitere Kommandos nachschauen, außer Punkt und Komma.

Außerdem habe ich mit Erika und Alexis wieder Deutsch geredet und auch etwas Neues über Japanisch gelernt. Es macht immer wieder Spaß sich auf Deutsch mit ihnen zu unterhalten.
Japanisch ist schon eine sehr simplizistisch Sprache, wodurch sie eben besonders schwer wird.