rictin.com

XPath shell script

Sat, 12 Apr 2014 23:00 +0000

I often want to quickly run an XPath expression against an XML file, e.g. when grep is not satisfactory. Commands I have tried usually don't even support "*:name" to ignore namespaces. I wrote a script based on xsltproc that ignores all namespaces, selects nodes matched, and prints an optionally customized expression for each node:

Usage: xpath FILE XPATH_EXPRESSION [XPATH_EXPRESSION]

Name spaces are ignored. For each node matched by the first
expression, the optional expression (default: ".") is applied.

Example: xpath file "/books/book[year=1983]" "concat(title,author)"

The output consists of two columns; the path to the matched node, and the value of the matched node (the 2nd expression). Example:

$ xpath pom.xml //version
/project/version        0.2
/project/build/plugins/plugin[1]/version        3.1
/project/build/plugins/plugin[2]/version        2.6

I also wrote a completion script to enable auto completion of XPath expressions, like:

$ xpath pom.xml /project/de[TAB][TAB]
/project/dependencies/  /project/developers/

The files: xpath.sh, xpath_completion.

Book.js

Thu, 20 Mar 2014 19:00 +0000

There are a lot of page flip libraries available, but I couldn't find a simple one that was compatible with old browsers and with a permissive software license, so I wrote one. Demo:

<div class="book pages_5">
	<img src="http://data.rictin.com/img/book.js/page0.jpg" alt="Page 0" />
</div>

Result:

Click on the images to turn page.

Code: book.js (BSD license)

Readable list operations

Fri, 15 Nov 2013 21:00 +0000

My attempt at simplifying common (java.util.)list operations such as:

	Lists.sort(list).ascendingBy().getFirstName()
	Lists.sort(list).nullsFirstDescendingBy().getFirstName()
	Lists.filter(list).accept("Erik").getFirstName()
	Lists.filter(list).rejectLessThan(18).getAge()
Long total =	Lists.sum(list).getCredit()
	Lists.search(list).forValue("Erik").getFirstName() != null
int minScore =	Lists.search(list).forSmallest().getScore()
int minScore =	Lists.min(list).getScore()
	Lists.forEach(list).setScore(0)
	Lists.select(fromList, toList).acceptGreaterThan(30).getAge()
	Lists.collect(persons, names).getName()
	Lists.groupBy(inputList, resultMap).getAge()

I think this has some very readable and easy to use code. Compare these (no pun intended. sort of. sorry.) different ways of sorting:

Pure Java:

Collections.sort(list, new Comparator<Person>() {
	public int compare(Person o1, Person o2) {
		if (null == ...
		return o1.getAge() - o2.getAge();
	}
});

Google Guava:

List<Person> sorted = Ordering.natural().nullsFirst().onResultOf(
		new Function<Person, Integer>() {
	public Integer apply(Person person) {
		return person.getAge();
	}
}).sortedCopy(list);

Lambdaj:

List<Person> sorted = Lambda.sort(list, Lambda.on(Person.class).getAge()); // nulls?

Mine:

Lists.sort(list).nullsFirstAscendingBy().getAge();

Google Guava and Lambdaj are much more powerful and feature complete, but I think mine is nice in specific cases.

Download: lists-0.2.jar (Java 5, source included, MPL2).

Javadoc

Known limitations:

Supports only one method call per element. WON'T work: sort(list).ascendingBy().getAdr().getCountry()
List cannot be empty
Final class elements won't work, but special support has been added for java.lang.String/Integer/Long/Double

Valgomat

Thu, 08 Aug 2013 23:59 +0000

Sorter utsagnene ved å dra boksene med musen slik at det du er mest enig med står øverst.

[Partiets] prinsipielle syn er at [...] ingen posisjoner [skal] gå i arv. Eventuell avvikling av kongedømmet må forankres i folkeavstemming

I et demokrati kan ikke viktige samfunnsposisjoner gå i arv. [Partiet] vil derfor avskaffe monarkiet og innføre republikk som styringsform i Norge.

[Partiet] slår ring om [...] monarkiet [...] og andre samlende nasjonale institusjoner fordi vi mener at disse er bærende i den norske identiteten

I dag tas for mange avgjørelser i internasjonale organisasjoner som er med på å befeste de rådende maktforholdene og forskjellene, som IMF, Verdensbanken, WTO og G20. At disse demokratiseres eller erstattes med et sterkere FN og mer demokratiske institusjoner

[Partiet] ønsker ikke en ytterligere liberalisering av verdenshandelen. Ikke alle sektorer egner seg for liberalisering gjennom internasjonale handelsavtaler. Vi vil støtte opp om interessene til de fattigste landene i WTO.

mekanismer som favoriserer rike land, må reduseres, og erstattes av et handelsregime som ivaretar utviklingslandenes interesser på en bedre måte. Norge bør aktivt bidra til å videreutvikle og styrke WTO. Systemet må reformeres, åpnes og demokratiseres.

Vi ønsker å begrense kommersiell finansiering av skoler [...]. Dagens privatskolelov dekker et tilbud med lange tradisjoner som et supplement til fellesskolen.

Velferdsstaten skal gi skole [...] av god kvalitet til alle. [...] Det er et offentlig ansvar å sørge for velferd for alle. Større innslag av private tilbydere av velferdstjenester bidrar til å undergrave dette [...]

Gode private skoler skal ikke møtes med frykt og forbud, men med en satsing på den offentlige skolen som gjør at den foretrekkes av de fleste.

Staten og andre offentlige institusjoner skal i toleransens navn være upartiske [...] staten må bidra til rause finansieringsordninger for tros- og livssynssamfunn. [Partiet] vil skille stat og kirke

kirke, religion og livssyn fortsatt bør være et offentlig anliggende. [Partiet] vil: - holde på en grunnlovsforankret folkekirke

[Partiet] vil derfor avskaffe statskirkeordningen og sikre at offentlige institusjoner er livssynsnøytrale, samtidig som de enkelte trossamfunn får frihet og mulighet til å drive sin virksomhet

det offentlige bør eie dominerende selskaper i markeder som er av avgjørende økonomisk betydning for landet. [...] Vi må [...] utvide offentlig eierskap.

Offentlig eierskap i næringslivet skaper lett rolleblanding og demokratiske problemer. Det skal derfor begrenses til vitale samfunnshensyn, som viktige kompetansemiljøer, grunnleggende infrastruktur og kontroll med energi- og naturressurser.

ønsker vi norsk privat eierskap med lokal forankring og statlig deleierskap der det trengs for å sikre nasjonalt eierskap. [...] Offentlig eierskap kan være hensiktsmessig i næringer og bedrifter med nasjonal, strategisk betydning.

[Partiet] er for et alleuropeisk samarbeid. [...] Forholdet til EU må demokratiseres slik at flere avgjørelser prøves i offentligheten. Dette krever en opprydning i de politiske prosessene i Stortinget, slik at en skiller ut de vesentlige politiske sakene fra andre mindre vesentlige. [...] et eventuelt norsk EU-medlemskap skal avgjøres etter folkeavstemning

Overnasjonale ordninger av denne art [EU] er et dobbelt onde; folkestyret svekkes og blir erstattet med en nyliberalistisk økonomisk orden [...] Utenfor EU står Norge langt friere til å kjempe for global rettferdighet [...]. [Partiet] er derfor motstandere av norsk EU-medlemskap og EØS-avtalen.

[Partiet] mener at Norge fortsatt skal stå utenfor EU, og til at vi vil si opp EØS-avtalen og erstatte den med handels- og samarbeidsavtaler med EU. Det skal ikke være EU som utformer regler for det norske samfunnet.

hoveddelen av norske klimakutt skal tas innenlands [...] Skogen er en viktig del av løsningen i klimapolitikken. Skogen i Norge tar opp CO2 tilsvarende halvparten av de norske klimautslippene.

alle samfunnsbeslutninger skal bygge på noen grunnleggende miljøprinsipper [...] Redusert forbruk i vår del av verden er en forutsetning for å løse mange miljøproblemer.

Forbrukerne må alltid kunne velge miljøvennlig. [...] [Partiet] arbeider for bruk av grønne skatter, omsettelige kvoter og andre markedsmekanismer for å få ned klimautslippene både nasjonalt og internasjonalt.

[Partiet] arbeider for utviklingsfremmende internasjonale handelsregler som ivaretar interessene til fattige folk i fattige land. Handelsreglene må gi utviklingslandene politisk handlingsrom, og ikke frata dem virkemidler som har vært sentrale i vår egen utvikling [...] Lokal matproduksjon må kunne skjermes mot å bli utkonkurrert av subsidierte landbruksprodukter fra rike land.

Norge må bli mindre opptatt av å beskytte seg mot internasjonal konkurranse og mer opptatt av å mestre sin rolle i en globalisert verden. Forenkling av regelverk og kamp mot handelshindringer, monopoler og maktkonsentrasjoner må til for å skape mer rettferdige spilleregler.

Importvernet for jordbruksvarer er avgjørende for å ha et norsk jordbruksvaremarked. Det høye norske lønns- og kostnadsnivået sammen med et utfordrende klima og krevende topografi bidrar til at prisene for norske landbruksprodukter er høyere enn i våre naboland. Det kan derfor ikke inngås nye handelsavtaler som svekker importvernet og rammevilkårene for den nasjonale landbruksproduksjonen.

det skal svært tungtveiende argumenter til for at kriger som ikke er sanksjonert i FNs sikkerhetsråd, kan rettferdiggjøres. Kun akutte humanitære kriser rettferdiggjør slike militære aksjoner. [...] Gjennom NATO er Norge i stand til å forsvare sine grunnleggende sikkerhetsinteresser, men også økonomiske og politiske tiltak må til for å forebygge globale trusler

Gjennom NATO-medlemskapet er Norge del av en militær strategi som bygger på bruk av atomvåpen og åpner for offensive militære operasjoner utenfor NATOs område. [...] [Partiet] er mot norsk medlemskap i NATO. [...] [Partiet] tar avstand fra all bruk av militærmakt som ikke er hjemlet i FN-pakten.

mener at Norge fortsatt skal basere sin forsvars- og sikkerhetspolitikk på medlemskapet i Nato og aktivt arbeid i FN. [...] Deltagelse i FN-ledede operasjoner må prioriteres. [...] ikke avgi militære styrker til internasjonale operasjoner utenfor Nato uten forankring i FN-pakten og et klart FN-mandat

Neste

Linux find

Thu, 20 Jun 2013 20:19 +0000

Here are a couple of find tricks I use repeatedly.

Search files

By declaring this function in e.g. ~/.bash_profile I can efficiently search files.

grip(){ find "${3-.}" -name '.[^.]*' -prune -o -name "${2:-*}" -type f -print0 | xargs -0 egrep ${@:4} "$1" ;}

I also set GREP_OPTIONS='--binary-files=without-match --color'

It skips all files and directories whose name starts with a dot. It works very much like wcgrep which I discovered later.

Usage: grip PATTERN [FILENAME_PATTERN] [DIRECTORY] [GREP_OPTIONS]...

Default FILENAME_PATTERN if not set or empty: *
Default DIRECTORY if not set: .

Examples:

`grip something`	Searches all files in current directory for "something"
`grip something \*.txt`	Searches .txt files
`grip something '*.txt' /some/path`	Searches .txt files in /some/path
`grip something '' . -m 1`	Searches all files in current directory for first occurrence of "something"

Stop descent on first finding

Useful for finding outermost index.htm or pom.xml files, for example.

find -type d -execdir test -e {}/"$1" \; -print -prune

Find deepest directories (no subdirectories)

find -type d ! -execdir sh -c "find '{}' -maxdepth 1 -mindepth 1 -type d -print -quit | grep -q ." \; -print

This became more complex than anticipated. Because -execdir relies on exit code and find returns 0 (OK) on no results, I had to use grep, and because of the pipe, I had to use sh -c.

Another, faster, way is:

find -type d | awk 'NR>1 && $0 !~ "^" prev "/" { print prev } { prev=$0 } END { print }'

Grep log files

Here is a small program I wrote for easier grepping logfiles by printing the timestamp even if the matched line doesn't have one (e.g. stack traces).

awk -v "pattern=$1" '{
  if (match($0, /^[0-9][^[:alpha:]]*/)) {
    dat=substr($0, 1, RLENGTH)
    grep("")
  } else
    grep(dat)
}
function grep(prefix) {
  if ($0 ~ pattern) print prefix $0
}'

I also created a similar logcat.sh (nice with pipe from tail -f).

Celestial sphere

Sat, 20 Apr 2013 20:00 +0000

You should be redirected here.

Epub2pdf

Wed, 06 Mar 2013 22:11 +0000

This program is for converting an EPUB file to a PDF file for printing.

The goal is to have the result resemble the EPUB book as closely as possible. Of course, because the EPUB format is not meant for printing, this is not an exact science. Nonetheless, all styles specified by the EPUB author are preserved, and additional sensible defaults are used.

Parameters such as page size and margins are easily configurable.

Page numbers and hyperlinks are not deemed important because links are not meaningful for printing and page numbers do not exist in EPUB electronic books.

Jar w/dependencies & source (14.5 MB)

Usage:

epub2pdf epubFileName pdfFileName [styleString|styleFileName]

Example styleString: "@page{ size: 21cm 30cm; margin: 10%; }"

Shaxpeare.com

Fri, 04 Jan 2013 23:11 +0000

I created a web site for William Shakespeare's plays of the 1623 First Folio,

www.shaxpeare.org

While there exist many Shakespeare resources online, most of them emphasize acts and speeches. This website on the other hand, focuses on pages and words. The goal is to represent the original printed book as closely as possible, spelling mistakes and errors included, but also to be accessible and useful (searchable).

The content on the site is mainly based on a modified "00ws110.txt" from Project Gutenberg. I use Apache Maven to build the site, and I wanted few design time and run time dependencies and a small client footprint.

Facebook book

Wed, 05 Sep 2012 19:00 +0000

This is how I made a facebook book.

First we need to get the data. We don't have to write an app, we just need an access token and a url. This can be obtained e.g. by browsing to Facebook's developer resources. After logging in you can find the following link:

https://graph.facebook.com/me/photos?access_token=xxx

"/me" could be substituted with a friend's username or profile id. Append "&limit=999" to get more at once. The data is returned in json:

{
   "data": [
      {
         "id": "10101470046392053",
         "from": {
            "name": "Richard H. Tingstad",
            "id": "1245408"
         },
         "picture": "https://fbcdn-photos-a.akamaihd.net/xxx_s.jpg",
         "source": "https://fbcdn-sphotos-a-a.akamaihd.net/xxx_n.jpg",
         "height": 540,
         "width": 720,
         "images": [
            {
               "height": 1536,
               "width": 2048,
               "source": "https://fbcdn-sphotos-a-a.akamaihd.net/xxx_n.jpg"
            }, ...

To turn this into a printable book I chose to use XSL-FO. This means that the data first has to be converted into xml. We get something like:

<data>
	<id>10101470046392053</id>
	<from>
		<name>Richard H. Tingstad</name>
		<id>1245408</id>
	</from>
	<picture>https://fbcdn-photos-a.akamaihd.net/xxx_s.jpg</picture>
	<source>https://fbcdn-sphotos-a-a.akamaihd.net/xxx_n.jpg</source>
	<height>540</height>
	<width>720</width>
	<images>
		<height>1536</height>
		<width>2048</width>
		<source>https://fbcdn-sphotos-a-a.akamaihd.net/xxx_n.jpg</source>
	</images>
	<images>
		<height>540</height>
		<width>720</width>
		<source>https://fbcdn-sphotos-a-a.akamaihd.net/xxx_n.jpg</source>
	</images> ...

There are many root "data" elements, so I added an all-surrounding "foo" tag. Next, I wanted to download the image files. Every photo has several sizes. To get the largest ones, I used the following xslt:

<xsl:for-each select="images">
	<xsl:variable name="w" select="number(width/text())"/>
	<xsl:if test="not(../images[width &gt; $w])">
		<xsl:value-of select="source"/>
		<xsl:text>
</xsl:text><!-- newline -->
	</xsl:if>
</xsl:for-each>

The XPath expression "/foo/data/images[1]" would probably work too, as they appear to be in the correct order. I piped the results to a text file, then downloaded them:

cat images.txt | while read a;do wget $a;done

Ironically, while I want high quality images now, I usually upload low resolution files myself. It's one of the jobs of my photo script (which has been heavily updated recently). Therefore, I manually replaced some of my pictures with high resolution files. The major step in the book exercise is the XSL transform. Feel free to download and use this file. Apache FOP can use this file to generate a PDF file.

To print the book, I ordered a photo album and uploaded each page as a jpg file:

convert -density 300 file.pdf files.jpg

Next time, I'll consider including QR codes in my book too, linking to the photos on facebook.com.

Gollnokkelen.com

Tue, 14 Aug 2012 19:40 +0000

Vi blogger om turer på www.gullnokkelen.com.

Caesar shift

Sat, 23 Jun 2012 19:00 +0000

Custom alphabet Caesar's cipher encryptor/decryptor.

[HTML Form]

Client-side printer friendly webpage

Wed, 16 May 2012 20:00 +0000

I will here show how the CSS3 pseudo selector :target can be used to display printer friendly versions of static web pages, as well as how to combine it with other methods.

Let's start with the following file(s) where we use CSS2's media type:

HTML:
<body>
<div>
	<div id="navigation" class="noprint">
		...
	</div>
	...
</div>
</body>

CSS:
@media print { .noprint { display: none; } }

This simple example hides all elements with class .noprint when the page is printed. We can expand the example with the CSS3 :target selector like so:

HTML:
<body>
<div id="print">
	<div id="navigation" class="noprint">
		...
		<a href="#print">Print</a>
	</div>
	...
</div>
</body>

CSS:
@media print { .noprint { display: none; } }
#print:target .noprint { display: none; }

"#print:target .noprint" is now activated by going to page.html#print. The advantage of this is that users get to see a print preview in their browsers - and no scripting was necessary, neither on the client nor on the server - purely CSS!

Let's help the user more and bring up the print dialogue box if (s)he goes directly to #print:

HTML:
<body onload="load()">
<div id="print">
	<div id="navigation" class="noprint">
		...
		<a href="#print">Print</a>
	</div>
	...
</div>
</body>

CSS:
@media print { .noprint { display: none; } }
#print:target .noprint { display: none; }

Javascript:
function load() {
	if (location.hash === '#print') window.print();
}

Finally, we can use Javascript to help out clients without CSS3 support:

HTML:
<body onload="load()">
<div id="print">
	<div id="navigation" class="noprint">
		...
		<a href="#print" onclick="printPage();return false;">Print</a>
	</div>
	...
</div>
</body>

CSS:
@media print { .noprint { display: none; } }
#print:target .noprint { display: none; }
.doprint .noprint { display: none; }

Javascript:
function load() {
	if (location.hash === '#print') printPage();
}
function printPage() {
	document.body.className += ' doprint';
	location.hash = '#print';
	window.print();
}

The beauty of this is that the print style sheet will work with CSS2, and the browser preview will work with CSS3 or Javascript.

Demo: Print page

P.S. Challenges:

The Back button may not function. Maybe a "back to full page" link should be provided (but not for @media print).
It may be tempting to merge the two last CSS rules, but this will break .doprint if e.g. Internet Explorer does not understand :target.
If printPage() is renamed print(), window.print() will be confused.
The id="print"-element has to contain all elements you want to hide or alter.

Update: Using CSS to display hidden content

Initially, I implemented a "back to full page" link basically by displaying this "hidden" element:

<p style="display:none" id="cancel"><a href="#">Cancel print preview</a></p>

A disadvantage of this is that the element is displayed when CSS is disabled. So I created a cooler trick, where CSS reveals the link! Here is the HTML:

<p><a href="#"><span class="reveal" title="Cancel print preview"></span></a></p>

The link here is hidden because the anchor tag has no content. I put the hidden text in the title attribute of a contained span tag. To display it, we can use the following simple CSS:

.reveal:after { content: attr(title); }

Here, CSS2 generates text content for the anchor tag (from the title attribute of the span tag), effectively revealing the link :-)

Chifre i Shakespeare Mysteriet

Tue, 17 Apr 2012 19:00 +0000

Boka Shakespeare Mysteriet av Søren Hauge handler om William Shakespeare, Francis Bacon og chifferteknikker benyttet på 15- og 1600-tallet, inkludert steganografiske metoder, altså skjult tekst.

Litt inn i kapittel 1 kom jeg over det jeg først trodde var typografifeil, hvor deler av teksten hadde en litt annen skrifttype enn resten. Etter hvert forstod jeg at forfatteren har tatt i bruk de omtalte kryptografiske metodene i egen bok.

Chifrene jeg oppdaget i boka er av typen Bacons bilitterære chiffer, hvor to skrifttyper brukes for å kode en beskjed.

Her er for eksempel chifferet på side 45:

han var en del af den litterære kreds

hanva	rende	lafde	nlitt	erære	kreds
baaab	aabbb	aaaaa	baabb	aabba	aabaa
S	H	A	U	G	E

Altså står det S. Hauge, forfatterens navn.

Sidetallene hvor disse chifrene opptrer er heller ikke tilfeldig valgt. Som boka forklarer kan man summere verdiene av bokstavene i et ord ved å sette A=1, B=2, osv. Navnet Bacon har f.eks. verdien 2+1+3+14+13=33. Side 53 og 157 har en prikk ved sidetallet.

Her er alle kodene jeg har funnet:

Side	Kode
33	Lys til din vej	33=Bacon
45	S. Hauge	45=?
72	K. Puggaard	Kirsten Puggaard har laget omslags-layout. 72=Puggaard
103	Hjertets sprog viser vej	103=Shakespeare
157	Til rosenkorsets visdom	157=Antall ord i To the Reader
277	Global Albada	Baklengs! 277=WilliamShakespeare+FrancisBacon

Side 107.

Organisten

Ekstrabonus: chiffer i boka Organisten av Erlend Loe og Petter Amundsen!

OrGan	isteN
babaa	aaaab
W	B

I denne boka er også sidetallet på side 399 endret til 993. Dette er en referanse til Shakespeares førstefolio hvor dette forekommer på den siste siden.

Hidden truth

Oppdatering 2013: Utfordring for leseren!

Bookmarkweb

Mon, 16 Apr 2012 22:00 +0000

Some months ago I wrote a GUI version of my Quix-like bookmarklet. It is still an open source JavaScript file, but a GUI allows for more interaction with the web page (e.g. the Regular expression find function). It's also simpler to use.

The bookmarklet:

Save the link (on the image/icon) or drag the icon to your toolbar.

A typical use case is selecting some text on a page, clicking the bookmarklet, and then selecting one of the actions to take; for example searching YouTube for the text, or tweeting the text and the URL of the page. Shift or ctrl click an action to open it in a new window or tab.

To try it out right now, just click it.

There are still some bugs, but I think it's usable.

Decoding the Ast Ritual

Tue, 27 Mar 2012 21:42 +0000

When I stumbled upon the page about the Ast Ritual at the Scottish Rite Masonic Museum and Library, I immediately wanted to know more and started decoding it.

Transcription

;Ast SC PE's L.

1. FM whce cm u ]oviz[ sa na PE ]gmz[ 2. fm ]gp[ e g f h st

jn t j ]mqz[ rsa ]pq[ lm 5 wt cm u hr o d ]mpq[ 8 o lrn o sbdu

ym psns nd mprv flsm n ysm 3 tn u r a ]vs[ n i prsm 4 i m os

tkn nd cptd mg brs nd swlf 7 hw d u no urslf ]pinz[ o b a n S

y hvg bn ftn trd ]ema[ nv dnd nd m wlng o b trd gn 3 wt mks

u a n 2 ym obn 1 hw shl i no u o b a n 4 y a ertn sn nkt drw

nd e prft stnp f ym ntnc 1 wt r sns 2 rt slgna, sltnzrh, nd

srlednprp . 3 pls o gv em e sn f n Pe . (rss nd gvs t) S hs tt ny

llsn . 4 t hs o e ytnp f ym obn tt i wd rthr hv ym | ct ssrca fm

re o re thn o rvl e stres f erf ysm nl ]qmp[ fly . 5 wt s a nkt

6 a crtn fndly r ]emz[ bry prg wrb ;a br yam no nthr n e krd

sa wl sa n e xl . 7 pls o gv ur lft dnh br e nkt nd prg f n pe

(ty rs nd gv t) A gvs B e prg. B sys wt s tt A i liah B i cncl

A. wt d u cncl B l stres f ns n ysm xet t shl b nt a urt nd l

]nmz[ fl br r wn e ydb f a tsj nd l ]pmq[ fl g f sch A wt s tt

cld B a prg A prg f wt B Pe A hs t a nm. B t hs A wl u pls

o gv t o em B i dd ut os rev t nthr en i os mpt t A hw wl u

dsps f t B i wl rettl r vhl t wh u A u wl pls o rettl t B . nigb

A u nigb f u pls (eton A gvs fast rettl Be ;e B poncs e drw) A

t s rt i grt u br * 1. whr wr u fast prpd o b md a n 2 n ym trh

S whr ;endy 4 n a moor djet o e g 5 hw wr u prpd 6 y bng

dvsd f l slrnm nd sltm nhr dkn nr clhd bt nr † ‡ wh a ¶ abt ym

ken n whh stn i ws cn ]vizh[ dctd o e dr f e g y e hd f a frnd

wm i ftrds fnd o b a br 7 hw dd u no t o b a dr bng ‡ 8 y fast

mtng wh rstnc nd ftrwds gng dmssn 9. hw gnd u dmssn 2 y ;i

dstnt skonk 3 wt ws sd o u fm wn 4 wo cms thr 9 ur nsr 2 n

roop dnlb cn ]amz[ ddt wo hs gnl bn dsrs f bng brt tuo f drkn

o xl o hv nd rev a trp f e rts nd bnfts f ts wRt g retd o § nd

ddctd o h st jn sa ynm brs nd swllf hv dn bfr em 1 wt mr wr

u skd 4 f i cm f ym wn fr wl nd crd f i ws dly nd trly prprd

yhtrw nd wl dflq f lwfl ega nd prpy dhcva fr 1 whh bng nsrd

u e vtrff i ws skd y wt rhtrf rt r bnft i xpcd o gn dmsn 7 ur

nsr 8 y bng a mn erf brn f gd rprt nd wl remd 9 wt flwd 2 i

ws rqsd o wt a tm wh ptnc ntl e wrt rm cd b nf ]qpt[ rmd f

ym rq ]azp[ st nd s nrs rtd . 1 wt ws hs nsr wn rtd 4 . lt mh

ntr . S hw dd u ntr 6. no e pt f a prhs nst ]enw[ rmt preg ym

dkn lft tsrb 5 whr wr u tn cn ]prz[ dctd 8 o e cnr f e g whr i

ws csd o lnk nd rev e tfnb f a ryrp (rm gvs ;i skonk 1 rs) i wl

thnk u o rpt e ryrp “ veh ]qmz[ thu dia” (es Mnr) 1. ftr revg

Translation

First Section Entered Apprentice's Lecture.

1. FROM whence came you ]oviz[ as an Entered Apprentice ]gmz[ 2. From ]gp[ the Lodge of Holy Saint

John at Je ]mqz[ rusa ]pq[ lem 5 What came you here to do ]mpq[ 8 To learn to subdue

my passions and improve myself in Masonry 3 Then you are a ]vs[ Mason I presume 4 I am so

taken and accepted among brothers and fellows 7 How do you know yourself ]pinz[ to be a Mason S

By having been often tried ]ema[ never denied and am willing to be tried again 3 What makes

you a Mason 2 My obligation 1 How shall I know you to be a Mason 4 By a certain signs token word

and the perfect points of my entrance 1 What are signs 2 Right angles, horizontals, and

perpendiculars . 3 Please to give me the signs of Mason Entered Apprentice . (rises and gives it) S Has that any

allusion . 4 It has to the penalty of my obligation that I would rather have my throat cut across from

ear to ear than to reveal the secrets of free Masonry unlaw ]qmp[ fully . 5 What is a token

6 A certain friendly or ]emz[ brotherly grip whereby one brother may know another in the dark

as well as in the light . 7 Please to give your left hand brother the token and grip of an Entered Apprentice

(ty rise and give it) A gives B the grip. B says what is that A I hail B I conceal

A. What do you conceal B All secrets of Masons in Masonry except it shall be to a true and law

]nmz[ ful brother or within the body of a just and law ]pmq[ ful Lodge of such A What is that

cld B A grip A Grip of what B Entered Apprentice A Has it a name. B It has A Will you please

to give it to me B I did not so receive it neither then I so impart it A How will you

dispose of it B I will letter or halve it with you A You will please to letter it B . Begin

A You begin if you please (note A gives first letter Be two B pronounces the word) A

It is right I greet you brother * 1. Where were you first prepared to be made a Mason 2 In my heart

S Where secondly 4 In a room adjacent to the Lodge 5 How were you prepared 6 By being

divested of all minerals and metals neither naked nor clothed barefoot nor shod hoodwinked with a cable-tow about my

neck in which situation I was con ]vizh[ ducted to the door of the Lodge by the hand of a friend

whom I afterward found to be a brother 7 How did you know it to be a door being hoodwinked 8 By first

meeting with resistance and afterwards gaining admission 9. How gained you admission 2 By three

distinct knocks 3 What was said to you from within 4 Who comes there 9 Your answer 2 A

poor blind can ]amz[ didate who has long been desirous of being brought out of darkness

to light to have and receive a part of the rights and benefits of this Worshipful Lodge erected to God and

dedicated to Holy Saint John as many brothers and fellows have done before me 1 What more were

you asked 4 If I came of my own free will and accord if I was duly and truly prepared

worthy and well qualified of lawful age and properly vouched for 1 Which being answered

you the affirmative I was asked by what further right or benefit I expected to again admission 7 Your

answer 8 By being a man free born of good repute and well recommended 9 What followed 2 I

was requested to wait a time with patience until the Worshipful Master could be inf ]qpt[ ormed of

my reque ]azp[ st and his answer returned . 1 What was his answer when returned 4 . Let him

enter . S How did you enter 6. On the point of a sharp inst ]enw[ rument pressing my

naked left breast 5 Where were you then con ]prz[ ducted 8 To the centre of the Lodge where I

was caused to kneel and receive the benefit of a prayer (Master gives three knocks 1 rs) I will

thank you to repeat the prayer “ veh ]qmz[ thu dia” (es Mnr) 1. ftr revg

The ritual is written in a kind of shorthand with some twists. The square brackets and letters between them look like null ciphers, and some words are not easy to guess. Words written in italic should be read backwards. I like the use of special symbols (¶=cable-tow, §=God), and the numbering system, which I interpret as ; followed by a vowel such that a=1, e=2, and i=3 (and o=4?). I don't understand why one word has a capital letter in the middle: wRt (worshipful).

It's also interesting to see how similar this shorthand from 1822 is to modern SMS language: "hw d u no urslf".

Rick Dangerous II dynamic font

Sat, 24 Mar 2012 13:49 +0000

My take on dynamic text replacement. I used the font of Rick Dangerous 2, but the technique is valid for any generated font image (see e.g. NFG Arcade Font Engine).

1. Image generation

A simple PHP script, rick-png.php, generates the image dynamically using a bitmap file of the font alphabet.

2. Text replacement

The following javascript replaces all text in h2 tags with generated images. I wanted the replacement to happen upon completion of generating and loading the image.

function rick(){
	var e=document.getElementsByTagName('h2');
	var a=[];
	for(var j=0; e.length>j; j++){
		var t=e[j].textContent;
		if(t==undefined)t=e[j].innerText;
		a[t]=e[j];
		var i=document.createElement('img');
		i.alt=t;
		i.onload=function(){
			a[this.alt].replaceChild(this,a[this.alt].firstChild);
		};
		i.src='http://www.gullnokkelen.com/rick-png.php?t='+t;
	}
}
window.onload=rick;

This method should be good for search engines and for unsupported clients, because the original text will be used.

P.S. In the code above, we depend on the a[] array and image alt attribute, but a perhaps cooler way is to use closures (which also works with multiple identical texts, unlike the former method):

function rick(){
	var e=document.getElementsByTagName('h2');
	for(var j=0; e.length>j; j++){
		var t=e[j].textContent;
		if(t==undefined)t=e[j].innerText;
		var i=document.createElement('img');
		i.alt=t;
		i.onload=clos(e[j]);
		i.src='http://www.gullnokkelen.com/rick-png.php?t='+t;
	}
	function clos(e){
		return function(){
			e.replaceChild(this,e.firstChild);
		};
	}
}
window.onload=rick;

P.P.S. Wouldn't it be cool if this CSS could work: h2{ content: url('http://www.gullnokkelen.com/rick-png.php?t=' attr(title)); }

Pigpen cipher on Mark Master Masons token

Tue, 20 Mar 2012 20:57 +0000

The pigpen cipher is fun because is it generally easy to understand, yet has several variations. Here are a couple I haven't seen deciphered on the internet, so I challenged myself to solve them.

Masons have these "pennies" or tokens that they give to candidates, some of which have ciphers on them.

The exact key to this particular cipher isn't reproduced much on the internet, but I finally found it:

The solution is: JOPPA- KEBRAOTH MARK.WELL

I gather from one of the very few Google results for JOPPA KEBRAOTH that these are some kind of freemason passwords.

Here is another one:

(www.marktokens.org.uk)

This one uses a different key, similar to a normal version, but backwards. I couldn't find an existing image, so I assembled this:

The solution is: THE MARK PROVINCE OF EAST LANCASHIRE

XSLT Content Management

Sat, 18 Feb 2012 22:20 +0000

I wanted this web site to be simpler and more portable, so I threw out Wordpress, PHP, and MySQL, and wrote my own code based on one XML and one XSLT file.

Here is an excerpt of the XML file:

<post>
	<date>2012-02-18T22:20:00</date>
	<url>xslt</url>
	<title>XSLT Content Management</title>
	<css>pre{ background-color: #EEEEEE }</css>
	<excerpt>I ported my web site from PHP to XML and Java.</excerpt>
	<tags>
		<tag>xslt</tag>
		<tag nolink="true">xsl</tag>
	</tags>
	<tns:content xmlns="http://www.w3.org/1999/xhtml">
		<p>I wanted this web site to be simpler and more portable,

The XSLT file creates several (X)HTML files using the xsl:result-document instruction. Where I needed server side dynamics, I used JSPX and Java.

Features:

Portable
Fast (no database)
Secure (static files)
Separates content from style and code
Tags
RSS feed
Links to next and previous post
"Hidden"/unlisted posts (pages), etc.
Support for custom CSS + JavaScript

To simplify building and error detection, I configured this XML pipeline using maven-antrun-plugin:

Validate content.xml
Transform with XSLT
Validate HTML files (XHTML 1.1 Strict)
Check well-formedness of JSPX files and feed.xml

Using Opera Mini as an anonymizer

Wed, 21 Sep 2011 20:25 +0000

Because Opera Mini fetches all content through a proxy server over an encrypted connection, you are guaranteed anonymity at the user end.

This prevents eavesdropping by your neighbor at least. The more recent versions of Opera Mini can be run in mobile emulators and look rather well on larger displays:

I then noticed two cool things: 1. MicroEmulator can run MIDlets from URLs using Java Web Start, and 2. When using the Opera Mini Demo, the client IP address is not sent in the X-Forwarded-For HTTP header. This is what gives us anonymity. It’s not as secure as Tor, but more practical. All you need is to click one link!

http://www.rictin.com/mini

(I recommend: Options > Select device > Resizable device)

(The link forwards to http://~~microemu.org/webstart/~~www.rictin.com/files/mini.jnlp where www.rictin.com/mini.(jad|jar) is the Opera Mini Demo)

P.S. Opera Mini does not provide end-to-end security. Also, we can’t control the HTTP headers or what Opera decides to do on its servers.

Join video files Nautilus script

Sat, 07 May 2011 21:23 +0000

The command to join video files with mencoder looks something like this:

mencoder -oac copy -ovc copy -o out.avi in1.avi in2.avi in3.avi

There are also other parameters one would want to set depending on the video file formats. To simplify this process of merging videos, download

Join video files

and place the file in ~/.gnome2/nautilus-scripts

Then, simply select and right click video files in Nautilus and choose Scripts -> Join video files

The script will try to set the correct options based on the video files. Files are joined in sequence according to their file names in alphabetical order. The output file is named using the name of the first video file; v1.mpg + ... => v1-joined.mpg As shown in the command above, the video and audio streams are copied, not re-encoded.

Zodiac tool

Fri, 22 Apr 2011 15:32 +0000

I decided to have a go at the unsolved "340 cipher" from the Zodiac killer.

After fruitless attempts using pen and paper I created a web TOOL.

There exists several similar tools already, but there are some differences. I use numbers instead of symbols to easier process the data. I think maybe the greatest advantage of my tool is the hover effect which affects all similar symbols on mouse-over. There is also a (rather primitive) "generate remaining" function which estimates unsolved symbols.

Unfortunately, I still haven't solved the cipher ;-)

Send custom HTML via Gmail from command line

Wed, 05 Jan 2011 20:35 +0000

Gmail allows composing and sending of HTML email, but it does not let you edit the source. Among other things, this irritates me because images with links get ugly borders. Luckily, Google also allows us to send emails from your account using other clients. Great, but composing HTML emails with inline/embedded images is still a hassle. Because I couldn't find anything better, I wrote a bash script that converts a webpage/HTML document to an email:

html2email.sh

The usage is simple: ./html2email.sh input.html output.txt [subject]

Then output.txt can be mailed by e.g. Gmail. A simple way of doing this is using sendemail:

sendemail -f you@gmail.com -t receiver@hotmail.com -s smtp.gmail.com:587 -o tls=yes -xu you@gmail.com -xp yourpassw -o message-format=raw -o message-file=output.txt

View the script

P.S. Edit output.txt to set To: field.

Script that prepares photos for Facebook

Tue, 04 May 2010 19:12 +0000

This is a follow-up of an earlier post. The topic remains the same; merging photo albums from different cameras while preserving chronological order, and removing potentially private metadata.

Like many others, I have noticed that Facebook increased its photo size from 604 to 720 pixels. At the same time, I noticed that the commands in my last post weren't perfect. I have tried to write an improved script:

photosort.sh

Usage: photosort.sh [OPTIONS] source_dir_1...[source_dir_N] target_dir

To find images deep in subfolders, just remove or edit "-maxdepth 1".

As before, I can recommend ImageMagick and Cygwin if you are a Windows user.

Shouldn't have to mention it, but: Use at your own risk.

Send SMS fra kommandolinjen

Sat, 13 Mar 2010 14:20 +0000

Oppdatering: Gulesider.no er endret så dette virker nok ikke lenger.

Gulesider.no lar norske mobilbrukere opprette konto og sende inntil 5 gratis SMS per dag fra siden sin (fra ditt eget mobilnummer!).

Jeg skrev et shell-script som sender fra denne siden uten at du trenger å bruke nettleseren. Dette er bl.a. nyttig hvis du ønsker å varsles av datamaskinen din ved ulike hendelser. Jeg synes også det forenkler manuell sending.

Last ned sms.sh og kjør f.eks.:

./sms.sh 98765432 Hei! Skal vi treffes i dag?

Lag en tekstfil, smsnums.txt, med 'nummer,kallenavn' per linje for å slippe å huske telefonnummer.

Jeg bruker Ubuntu Linux, men hvis du vil kjøre scriptet i Windows, er det bare å laste ned Cygwin og wget.

Ubiquitous Web OS Command Line

Sat, 06 Feb 2010 22:21 +0000

Update: I wrote a similar bookmarklet with a GUI.

The last couple of years, I have been wondering who my "online profile" will belong to. First, applications started moving from the PC to the web, such as email and bookmarking. Then browsers started syncing bookmarks and customizations across different PCs and mobile devices.

All of these developments are really exciting, but I am not totally satisfied. For one thing, you have to choose and stick to one particular browser. Also, they do not sync everything that I would like (e.g. Greasemonkey scripts). A blog post about alternatives to Mozilla's Ubiquity introduced me to some great projects, and inspired me to start using Quix. Quix lets you access your bookmarks, custom searches, and other tools, from one place. It works across all browsers and doesn't need to be installed!

Not long after, however, I decided to not depend on Quix at all, and write a free, open version. While Quix lets you write your own command file, I also let you modify and store the code. This way, you don't have to worry about Quix crashing, going bankrupt, or eavesdropping on your traffic. You just have to host one JavaScript file (or use mine) and remember its address.

Clicking the following image invokes the script:

Drag it to your browser's toolbar for convenience. Alter the web address in the link if you want to use your own copy or modified version. The file: WebOsCmd.js

Save file by URL

Sun, 17 Jan 2010 13:49 +0000

Oftentimes, I have a URL to a file I want to download, but no simple way of downloading it. The problem is that some files (e.g. videos) open in the browser instead of giving you the "Save As..." dialog box.

LinkSave.net provides a text box where you can paste the URL and generate a link you can right click and "Save As...". But do you want LinkSave to know all the links you save?

I created a similar link generator where your URLs are not logged. The link is created on the client side using JavaScript, and you can view the source code to be sure. Easier than writing a new HTML file with the URL every time you want to save a file. Of course, wget is an alternative.

Linksave

Sun, 17 Jan 2010 00:00 +0000

Similar to LinkSave.net, but guaranteed no logging.

[HTML Form]

Paste the URL of the file you wish to save and click "Link".
Right click the link and choose "Save Target As..." or "Save Link As..."

Link will appear here

Codes in Shakespeare

Fri, 25 Dec 2009 21:29 +0000

In the Norwegian book, Organisten (Dan/Swe: Ekspedition/Expedition Shakespeare), Petter Amundsen presents several ciphers in the works of William Shakespeare. A newer English e-book and film have also been produced. I will here try to investigate some of the alleged ciphers.

Contents

Bacon acrostics

Rosy Cross

Star map

Comparing to quartos

Miscellaneous

Oak Island

Conclusion

Appendix - images

Bacon acrostics

To the Reader has a w which should be capitalized and the three first outermost letters spell TWO. On page 2, another TwO appears, next to "F BACon".

This Figure, that thou here seest put,

It was for gentle Shakespeare cut;

Wherein the Grauer had a strife

with Nature, to out-doo the life :

O,could he but haue dravvne his vvit

As vvell in brasse, as he hath hit

Can such acrostics be found anywhere? A search for BACon in the First Folio yields only two hits:

$ tr '\n' = < 00ws110.txt |egrep -io '=B[^=]*=+A[^=]*=+Con[^=]*' |tr = '\n'

Begun to tell me what I am, but stopt
And left me to a bootelesse Inquisition,
Concluding, stay: not yet

Banish'd this fraile sepulchre of our flesh,
As now our flesh is banish'd from this Land.
Confesse thy Treasons, ere thou flye this Realme,

Other simple acrostics are also rare: TwO (1), BACON (0), FBACO (0), OCABF (1), BACO (1), OCABF (0), ConAB (2), FBAC (1), 'O w T' (1).

I do not believe that all acrostics are intended ciphers. However, I think that some of my findings are interesting. I think it's peculiar that a BACon is on page 2 of Comedies when I found BACon reversed on page 2 of Tragedies, especially knowing that only one other BACon (without the preceding "F ") and one other reversed BACon (indented) exist. I also think it's funny that the only FBAC and OwT I found are on the same page.

The three parts of the First Folio are about 300 pages long, so the probability of no "page number collisions" (assuming random distribution of the "BACon"s) is about 1 * 897/900 * 894/900 * 891/900 ~= 98.01%.

Rosy Cross

This page is mentioned in epilogue 4 of Organisten (not in the first edition of the book), but not the circled parts. Petter Amundsen found "R-O-s-y C" between Rossi and crosse after I told him about my discovery of fame and confesse. It's at a 53 degree angle of course! It's discussed further in the e-book.

The e-book says the following about the other Rosie Cross triange: "a hyphened adjective with a peculiar feature: “un-Sunn’d”, the strangeness being the capital S."

I checked and found that it's not unique, but admittedly unusual:

$ grep '[uv]n-[A-Z]' 00ws110.txt
God saue King Henry, vn-King'd Richard sayes,
And therefore Ile vn-Crowne him, er't be long.
As Chaste, as vn-Sunn'd Snow. Oh, all the Diuels!
Appeare vn-Kinglike

I also verified another e-book quote: "A brief inspection of the complete Cymbeline uncovers no similar groups of words with capital letters C.R.C."

$ egrep -o 'C\w*\W+R\w*\W+C\w*' 00ws110.txt
Cedar, Royall Cymbeline

Actually, the word group in Cymbeline is the only C.R.C. combination in all of the plays.

Star map

I'm convinced that Boötes and Wain (and Swan) are hidden in the Shake-speare productions, but do they form the geometric symbols that Petter Amundsen claim? To be a perfect match, the stars should fit the following diagram:

The distance between Regulus (Leo) and Deneb (Cygnus) is 119.47700 degrees, the distance between Arcturus (Boötes) and Capella is 103.03220 degrees. All stars are the brightest of their constellation. We get a proportion of 119.47700/103.03220=1.15961, which is 99.4% of the expected 1.16667 (7/6). The length and width match, but is the figure straight? CapellaDeneb/ArcturusDeneb = 78.18639/80.85666 = 96.7%. And finally, do the 6 and 7 units long lines cross at the right place? RegulusDeneb and ArcturusCapella intersect at about +54°51'11", 11h45'14", and IntersectionToRegulus = 46.9390° and IntersectionToDeneb = 72.5801°, which gives us IntersectionToRegulus/IntersectionToDeneb = 0.6467 = 86.2% of 3/4.

The star map is partly based on acrostics like the one above. This one is on page 37 of The Two Gentlemen of Verona, the first page to be mislabeled as The Merry Wives of Windsor. I think it's curious that I find another star on the next, and final page of the play, which is also mislabeled.

The two stars Tania Australis and Tania Borealis mark the left hind foot of Ursa Major.

The page number 169 should be special to the Rosicrucians because 16=4^², 9=3^², and 169=13^², which is why I wonder about two possible stars in 3 Henry VI:

Tejat and Almeisan (a.k.a. Alhena) are the brightest stars in Gemini after Castor and Pollux (son of Cygnus, the swan). Tejat is the left "back foot" or heel of Castor, and Almeisan is the left foot of Pollux.

None of these stars fit the existing star map.

Comparing to quartos

I thought it a good idea to compare the First Folio to the previous quarto and had a look at 1 Henry IV.

My first observation is that the two are very similar, only a few words are changed, and mostly the same ones. Many & have been replaced by and, and many expressions such as by the Lord and Zounds have been removed. The FBACO acrostic spans two pages in Q5, which I think discredits it a bit. The line between the Lords is also uncertain, as I see no obvious signs of manipulation of their positions. The removal of the other Lords could be a good sign, but since they are all "by the Lord", it doesn't look as good. Much the same goes for page 53, but Bottes/BoOtes and Boots/Bootes seem very plausible.

Miscellaneous

Geometric patterns and shapes are difficult to evaluate because of their subjective nature. One that I think can be measured more objectively is the five Lords in a row on page 131 of Histories. The page has 12 Lords in total, 66 lines and 82 characters per line in average. For 3 randomly placed Lords to be aligned between two others, we have approximately (64/66)*(63/66)*(62/66)*1/82^³*nCr(10,3)=0.000189..., i.e. 0.02% probability of coincidence. Unfortunately, that number is neither large enough nor small enough to settle any real doubt on its own.

Finally, an acrostic I have found, which could be nothing, is TAROT on the final page of All's Well That Ends Well.

Oak Island

One of the claims I am hesitant to accept immediately is that there are references to Oak Island. If one were to point to Oak Island, however, I'll admit that the Swan is the star constellation closest to the mark:

(Click image for details)

Conclusion

I have only focused on a few of the discoveries here, notably those I've been unsure about or contibuted to. Personally, I am convinced that Rosy Cross, Boötes, and Bacon ciphers exist in the publications, but I'm not sure what they signify.

You may also be interested in my Summary of most convincing Bacon ciphers in Shakespeare.

Appendix - images

Boötes in Antony and Cleopatra

Page 346 of The Tragedy of Antony and Cleopatra contains the only TwO I have found beyond page 2 of The Tempest, and the only FBAC in the Folio. The OwT is part of "BOwTHS" (=Boötes).

BACon on page 2 of Tragedies

I have found only two Con-A-B, and it's fitting that one of them is on page 2, mirroring the B-A-Con on page 2 of Comedies. Also a possible BOWTIs anagram. The other instance is indented and prefixed with Fra.

37/53 degrees of the 3-4-5 triangle in 1 Henry IV

This page and its acrostic ciphers are mentioned in Organisten, but the line was not. It passes through the page's only two "Lord"s, intersects the upper left corner, and has a 37 degree angle. I'm still not completely sure whether the finds on this page is accidental or not.

BACO on page 281

The most promising BACO I've found is on page 281 (prime), The Winters Tale. I'm guessing the others are coincidences (Comedies p. 21 and Tragedies p. 127).

The final BACon is on Histories p. 27 (Richard II):

Hvilket Flax-lodd er best?

Tue, 07 Apr 2009 17:07 +0000

I disse førjulstider er det mange kalendere og lodd å sette seg inn i. Men hvilke gir egentlig de beste vinnersjansene? La oss ta for oss Flax-loddene. På baksiden står vinnersannsynligheten:

Julekalender:	50,0 % (1:2)
Million:	25,5 % (1:3,92)
Griseflax:	20,6 % (1:4,84)

Ser jo greit ut, kalenderen er best. Eller er den det? På kalenderen kan du vinne maksimum kr 500 000, mens du kan vinne opptil kr 1 000 000 på "Flax million". Dessuten er kalenderen dyrere. I tillegg kalkulerer Norsk Tipping selv de minste premier som "win". Altså, hvis du skraper frem 20 kroner på et lodd som kostet 50, så har du "vunnet". Følgende sannsynligheter kan være interessante:

	"Vinnersannsynlighet"	Sannsynlighet for ikke tap	Sannsynlighet for lønnsom gevinst
Julekalender:	50,0 %	20,0 %	10,0%
Million:	25,5 %	25.5 %	9,7%
Griseflax:	20,6 %	20,6 %	11,2 %

Som vi ser er forskjellige typer lodd best på forskjellige målsetninger. Vi kan beregne forventningsverdien for å finne ut hvor mye loddene faktisk koster oss (i snitt):

	Forventet gevinst	Pris	Forventningsverdi per krone
Julekalender:	kr 25,85	kr 50	kr 0,517
Million:	kr 13,75	kr 25	kr 0,55
Griseflax:	kr 5,10	kr 10	kr 0,51

Hvis du kun er interessert i å tape minst mulig penger bør du altså velge Flax million. Hvis en er interessert i sannsynligheten for premier av en viss størrelsesorden kan man se på denne oversikten:

Julekalenderen ser ut til å ligge høyest hele veien, bortsett fra på >= 250, hvor Flax million gjør det litt bedre (men husk at loddene ikke koster like mye). Også på >= 1 000 000 vinner selvsagt Flax million.

Hvis man snevrer seg inn mot store premier, kan man også følge med på kunngjøringer om gevinster og oppdatere sannsynligheten deretter.

Beregningene finner du i dette regnearket.

IMG Mosaic Generator

Tue, 24 Nov 2009 19:44 +0000

A simple open source photomosaic creator (with support for multithreading) written in Java.

For more technical information, visit Google code, where the code is hosted.

To use, download IMGMosaicGenerator.jar, and run something like:

java -jar IMGMosaicGenerator.jar bigpicture.jpg 100 75 mytilesdirectory

This will look for image files in mytilesdirectory. Every image file is then resized to resolution 100x75 and matched against every position of bigpicture.jpg:

The preview window is updated every time a tile is a "best match so far". Window can be resized:

Two versions are created (one normal and one "optimally dithered"):

The second version can be obtained when aborting the process by clicking in the window and invoking the following dialog box:

Output is in HTML. To save as image, Firefox add-on, Abduction! can be used.

Preparing photos for Facebook upload

Sat, 17 Oct 2009 15:23 +0000

Update: Improved version in a new post of mine.

A few tips for using Facebook and/or other services. You will need ImageMagick and Linux or Cygwin.

Order images correctly

Copying images to your computer, or rotating them, may ruin the files' time (of last modification). If your album contains images from several cameras, the file names can't be used to sort the images either. The following command renames the files to the second they were taken (see Exif).

c=0;for i in *.jpg;do t=`identify -verbose $i |egrep 'Exif:DateTime:'|sed 's/[^0-9]//g'`; mv $i "$t-$c.jpg";c=$[ $c + 1 ];echo $i;done

Remove excess information

Be sure you don't share more than you mean to! Also, saves you a couple of bytes to upload. (Copy files to a new folder first.)

for i in *.jpg;do mogrify -strip -resize 604x604 $i;echo $i;done

Of course, for some services, keeping the metadata could be useful.

Anonymize file names

Finally, you may want to rename the album file names:

c=0;for i in *.jpg;do f=`printf "%04u" "$c"`;mv $i $f.jpg;c=$[ $c + 1 ];done

Punktliste mot datalagringsdirektivet

Wed, 05 Aug 2009 20:28 +0000

En oppsummering av problemene med datalagringsdirektivet slik jeg ser det:

Alle lovlydige mister sin anonymitet
Kan misbrukes i fremtiden (ved maktskifte)
Kan misbrukes hvis uvedkommende får tilgang til informasjonen
Lett å senere "utvide" til verre formål ("slippery slope")
Gjør det vanskeligere med whistle blowing
Ett skritt i retning av at man må bevise sin uskyld
Kriminelle kan lett unngå overvåkningen
Lite igjen i forhold til prisen?

Dette er selvsagt litt forenklet, men det må en punktliste nødvendigvis være. Finnes det flere punkter? Skriv gjerne en kommentar.

Tekst fra carlchristian.net (Creative Commons Attribution-Share Alike 3.0 Norway License):

Personvern er en grunnleggende verdi i et demokrati. Personvernet innebærer en rett til å være i fred fra andre, men også en rett til å ha kontroll over opplysninger om seg selv, særlig opplysninger som oppleves som personlige. Etter EMK artikkel 8 er personvern ansett som en menneskerettighet.

Med en mulig norsk implementering av Datalagringsdirektivet (direktiv 2006/24/EF), som pålegger tele- og nettselskap å lagre trafikkdata om borgernes elektroniske kommunikasjon (e-post, sms, telefon, internett) i inntil to år, vil nordmenns personvern bli krenket på det groveste.

Datalagringsdirektivet ble vedtatt av EU 15. mars 2006, men fremdeles har den norske regjeringen ikke offisielt tatt stilling til om direktivet skal gjøre til norsk lov eller ikke. Gjennom EØS-avtalen har Norge en reservasjonsrett. Denne har aldri før blitt brukt, men så har man heller aldri stått overfor et direktiv som representerer en så stor trussel mot demokratiets grunnleggende verdier som det datalagringsdirektivet gjør.

Morse Code MIDlet

Sat, 20 Jun 2009 13:20 +0000

Inspired by the flashlight midlet, Torch.

The midlet sends morse code signals, it does not receive.

Allows you to communicate accross distances where your voice and body is not enough and your phone is otherwise useless.

For example if you're in a dark and noisy environment and you don't have the recipient's phone number or there is no network coverage.

A life saver! ;)

Download: http://www.rictin.com/m.jad. If that doesn't automatically work, try to also download http://www.rictin.com/m.jar. Source code.

FINN Greasemonkey script

Thu, 09 Apr 2009 15:45 +0000

Update: FINN.no har endret mye på koden sin siden dette ble laget så det virker sannsynligvis ikke lenger.

Jeg har flere ganger brukt FINN eiendom, og synes stort sett det fungerer godt. Jeg liker at man kan avgrense søk såpass fritt (antall soverom, størrelse, etasje, osv.). Jeg har derimot syntes at det kan være litt vanskelig å vurdere boligers beliggenhet, da område/sted kan være litt generelle og grensene kanskje ikke går der man ønsker. Man måtte til dels vurdere hver enkelt boligs adresse. Jeg syntes det var veldig kult at boliganonsene ble plassert i kart etter hvert. Det har imidlertid fortsatt noen svakheter. Et av problemene synes jeg er at kartet ikke gir en rask oversikt over boligenes beliggenhet i forhold til noe, f.eks. trikkestasjoner. Jeg mener det også er svært aktuelt å vite en boligs avstand i reisetid til jobb eller skole.

Mitt Greasemonkey-skript forsøker å vise boligers avstand til en oppgitt adresse på en rask og enkel måte. Dette gjøres ved å legge på kjøreavstand (kilometer og minutter) til en oppgitt adresse på hver oppføring i resultatlisten. For å gjøre dette mer nyttig kan man sortere resultlisten etter kjøreavstand.

Man kan også sortere etter kvadratmeterpris, totalpris (inkl. fellesgjeld) eller standard (ingen sortering). Dersom skriptet ikke finner variabelen det sorteres på for en bolig (f.eks. kjøreavstand eller kvadratmeterpris) blir boligen vist sist.

Også på objektsiden vises kvadratmeterpris og kjøreavstand.

Hvis man klikker på kjøreavstand-linken får man opp kart med kjøreanvisning mellom de to adressene fra Google maps. Kjøreavstand vil dessverre ikke virke i Opera, men man kan fortsatt sortere etter totalpris og kvadratmeterpris.

Det er enkelt å endre koden for å få gåavstand i stedet for kjøreavstand eller å endre retningen (som har noe å si pga. enveiskjørte gater o.l.).

Hvis flere elementer i HTML-koden på finn.no hadde hatt id, ville det vært mye enklere å lage dette skriptet!

Online Excel Translator

Tue, 07 Apr 2009 17:07 +0000

Currently not working

Translate Microsoft Excel formulas/function names between several languages.

[HTML Form]

P.S. Danish is the least supported language (but most similar to Norwegian).

Feel free to contact me with suggestions etc.

Sources used:

http://cherbe.free.fr/traduc_fonctions_xl97.html / http://dolf.trieschnigg.nl/excel/excel.html

http://www.probabilityof.com/excel.shtml#30

http://www.sslug.dk/locale/kenneth/dom/apps/gnumeric-hacks/intlfuncn.html

http://www.piuha.fi/excel-function-name-translation/index.php?page=english-polish.html

Excel sort

Tue, 07 Apr 2009 17:07 +0000

I recently created an explanation of how to automatically sort a list of data in Excel.

I wanted to present the technique as simply and effectively as possible. Pages like MrExcel use jpg images to show screenshots of Excel sheets. This works nicely, but I wanted something more usable and accessible. HTML tables are also frequently used around the internet, but often rather primitively.

I used a simple HTML table with border="1" and some cellpadding as a basis, and added CSS on top. I found that I needed the HTML border and cellpadding for the table to be readable without CSS. The CSS has two functions; to mimic the Excel look and to use colors for highlighting.

The HTML+CSS approach has both advantages and disadvantages compared to screenshots. Disadvantage: Difficult to show colored squares that span several cells, such as:

Advantages: Can show the formula of more than one cell at a time. The text can be selected and copied, etc. Instead of cell-spaning rectangles, I use different colors for the cell contents.

I wanted it to be possible to copy the whole table and paste it into Excel or OpenOffice. You can, but you have to delete the first column afterwards.

I ended up with two style sheets (Creative Commons license):

I am also wondering if I can do something more with the Excel functions. I could link to the function reference, like ~~VLOOKUP~~(dead link). ~~Or perhaps I should link to my translation page; VLOOKUP~~(defunct). The question is what language to translate to. Perhaps we could use Accept-Language from the HTTP header. For now, I just put a <span class="msexcel"> around the formulas to make them easily available to machine interpretation (e.g. Greasemonkey scripts).

~~After writing the page, I found a video presentation (Flash) of the same technique. They do not mention the +COUNTIF function however.~~ (dead link)

Excel dynamic sort

Tue, 07 Apr 2009 00:00 +0000

Sorting data in-place in MS Excel is easy. This technique however will sort a table without changing the original data. This means you can have several referring tables that will automatically update when the original data is modified. Without macros.

~~P.S. Use my Online Excel Translator if you want the functions in another language.~~

This is our initial table:

	B	C
1	Name	Score
2	Mallory	50
3	Bob	120
4	Eve	10
5	Alice	100

Let's sort by score. Fist we need to add a column to the left of the data, with the "RANK" of each row. We have to add +COUNTIF(C$2:C4;C5) if the sorted variable (score) is not necessarily uniqe.

	A	B	C
1		Name	Score
2	3	Mallory	50
3	1	Bob	120
4	4	Eve	10
5	=RANK(C5;C$2:C$5)	Alice	100

Then we add a column with the desired order of data, usually 1,2,3,.. or perhaps a descending order.

	A	B	C	E	F	G
1		Name	Score	Desired order
2	3	Mallory	50	1	Bob	120
3	1	Bob	120	2	Alice	100
4	4	Eve	10	3	Mallory	50
5	2	Alice	100	4	=VLOOKUP(E5;A$2:C$5;2;FALSE)	=VLOOKUP(E5;A$2:C$5;3;FALSE)

That's it!

Now, if score is updated in column C, the sorted list to the right will automatically update. Especially convenient if you want to sort the same data in different ways.

JavaScript: RegExp Find

Sat, 13 Dec 2008 10:51 +0000

This script is exactly what its name suggests: RegExp Find.

Works much like regular find in browsers, but is JavaScript and supports regular expressions.

Bookmarklet: Regex find (Drag to your bookmarks/toolbar)

Was previously a Greasemonkey/user script; ~~One can press (default) Ctrl + Shift + F to Find, Ctrl + Shift + N jumps to the Next match. Tested in Firefox and Opera.~~

~~Related, but different: Beef up the Find command in Firefox - "Greasemonkey script to highlight search entries relative to nearby content."~~

Resources: Bookmarkleter

User JavaScript: Direct-Link

Wed, 17 Oct 2007 08:38 +0000

This Greasemonkey script finds URIs in links on web pages and creates a direct link to them.

For example, if a page has a link to http://site1.com?u=site2.com, a link that has caption ^ and points to http://site2.com/ is created next to the original link.

Script file.

Version 0.22+:

Small syntactic improvements to URL regexp.
Better readability of URL regexp.
Handles Base64 encoded URIs.
Recursive check of in-URL URLs.

You can discuss this script over at userscripts.org.