I couldn’t find a comprehensive guide covering all the necessary components to get this working, so I decided to write one. However, this post is not a step-by-step tutorial; it assumes familiarity with FluentBit, AWS and OpenSearch Serverless configuration options.

Networking

I kept the dashboard publicly accessible (authenticated via the AWS web console session) and created an OpenSearch Serverless network policy with a managed VPC endpoint for API access. This enabled an internal AWS network connection between the OpenSearch API and our ECS Fargate containers.

For this setup to work, the VPC endpoint requires a security group that allows inbound HTTPS access from the ECS container’s security group.

IAM

The role under which FluentBit runs (in our case, task_role_arn in the ECS task definition), requires the following permission:

 1{
 2    "Version": "2012-10-17",
 3    "Statement": [
 4        {
 5            "Sid": "VisualEditor0",
 6            "Effect": "Allow",
 7            "Action": "aoss:APIAccessAll",
 8            "Resource": "arn:aws:aoss:<region>:<account-id>:collection/<collection-id>"
 9        }
10    ]
11}

This might raise some concerns, but it only grants the role access to data plane APIs for a specified collection in the account. More details here.

The actual access control is enforced through the OpenSearch Serverless data access policy, where we need to define the following permissions for the role and collection above:

aoss:CreateIndex
aoss:UpdateIndex
aoss:WriteDocument

Create the index

1. Visit OpenSearch Dashboards URL for the collection. The URL follows this format: https://dashboards.<region>.aoss.amazonaws.com/_login/?collectionId=<collection-id>.
2. In OpenSearch Dashboards, open the left navigation pane and select Dev Tools.
3. Create an index using the query: PUT <index-name>

FluentBit

Add the following output configuration as per the FluentBit documentation:

[OUTPUT]
    Name  opensearch
    Match *
    Host <collection-id>.<region>.aoss.amazonaws.com
    Port 443
    Index <index-name>
    Trace_Error On
    Trace_Output Off
    AWS_Auth On
    AWS_Region ${AWS_REGION}
    AWS_Service_Name aoss
    tls On
    Suppress_Type_Name On

Ensure you can read and tail FluentBit engine logs to monitor for delivery errors.

Verification

In OpenSearch Dev Tools, check the index information using the query: GET _cat/indices/<index-name>?v

Once the document count increases and no errors appear in FluentBit, proceed to create an index pattern in the Stack Management menu to explore your data in Discover.

Results

As an interesting data point, with approximately 650 logs per second and an average log size of ~6KB, indexing scaled to ~10 OCUs on average. Search capacity remained mostly at 1 OCU, peaking at 5 OCUs during active searches.

We decided not to proceed with this project due to its pricing and challenges in handling schema-less JSON logs, particularly those with high cardinality and varying data types. While it was roughly half the cost of Datadog, in our current context, the trade-offs were not justified.

After seven years, I decided to write a blog post again, only to find that my old toolchain hadn’t aged well. My blog was built on Jekyll, specifically the GitHub Pages gem. Over time, as I continually updated my local Ruby installation, I had to update the blog’s gems as well. After falling down a rabbit hole of troubleshooting gem and theme updates, I realized that GitHub Pages doesn’t keep up with Jekyll either. I could have made it work with Jekyll alone, but it felt like the right time to switch to something leaner and less likely to break over time.

I ultimately chose Hugo. It’s fast, feature-rich, and doesn’t require Ruby, a JavaScript runtime, or any extra dependencies.

Theme

My previous theme was still functional, but over the years, I became increasingly frustrated with the modern web. Inspired by the Bear blogging platform and its no-nonsense philosophy, I decided to strip away external fonts, analytics, and the unnecessary fixed sidebar.

I also wanted a theme that respects the system’s light/dark mode settings for better readability. I settled on hugo-bearcub. It offers more than I need, but I had some issues with the more minimal hugo-bearblog. I might revisit it later to simplify things further.

Posts

In Jekyll, I had a custom permalink configuration:

1permalink: /:year/:month/:title.html

Hugo uses a different default permalink structure, but I wanted to maintain my old links. While I could have used an alias (redirect) in the front matter, I decided against it—switching tools shouldn’t force me to change permalinks. If I migrate to another static site generator in the future, I’d have to juggle multiple URL schemes, creating a convoluted mess of redirects.

Instead, I configured Hugo like this:

 1# [...]
 2# Displaying the .html extension is considered "ugly" by some
 3# Note: Order matters! This should be defined at the top, above `menus` and `permalinks`
 4uglyURLs = true
 5
 6[permalinks]
 7  posts = '/:year/:month/:slug'
 8
 9[outputFormats]
10  [outputFormats.RSS]
11    mediatype = "application/rss"
12    baseName = "atom" # Preserves the `atom.xml` name from Jekyll for the RSS feed

In Hugo, embedding raw HTML in Markdown isn’t allowed (or recommended), so for some posts, I renamed file extensions to .html to render them as-is. For others, I rewrote snippets in Markdown or used shortcodes.

Speaking of shortcodes, Hugo only interprets variables and functions in templates, so I had to use them for handling assets in page bundles:

• Displaying images using the imgproc shortcode
• Linking to a resource within a post, such as a full-size image or a downloadable file
  See code

   1{{- /*
   2  Renders an absolute path to a page bundle or global resource for use in an <a href> attribute.
   3
   4  @returns template.HTML
   5
   6  @example 
   7
   8*/}}
   9{{- with $.Get "path" }}
  10  {{- with $i := or ($.Page.Resources.Get .) (resources.Get .) }}
  11    {{ .RelPermalink }}
  12  {{- end }}
  13{{- end }}

  This workaround seems necessary because of my custom permalink structure.

While a bit more complex than I’d like, this approach keeps post assets self-contained, avoiding reliance on external storage or a CDN.

Hosting

I’m grateful for the work GitHub has put into GitHub Pages—it served me well for many years. For now, I’ll continue using it for building and hosting this site.

While I was focusing on my personal development in the past year I kind of accepted that it is OK not having a professionally challenging job. However this can be a dangerous path, with a trap one can easily fall into: if there are no challenges, it must be because we already know everything!

Don’t get me wrong I was still learning but that happened outside of daily work. I see it more clearly now that I could improve much faster if work hours would be aligned with my professional growth as well.

With this in mind I’ve decided to look for new opportunities. I started my research to find companies I want to work for. It was a crucial moment when I realized while it’s useful, if I want to be honest with myself, I am procrastinating on sending the applications. I wanted to know more about my fears and fight them. After some self-introspection it turned out that I’m afraid of being inadequate and being a failure in general.

On an absolute scale which step of the career ladder am I on? It feels like I see some steps below me and the next one or two above me but I don’t know where this step exactly is in the industry.

If I wanted to sound scientific I would ask where the line is between the impostor syndrome and the Dunning–Kruger effect.

I am convinced that by going through some interviews again will help me cope with these feelings. And I want to gain some insights into where I must improve.

My biggest challenge now is the first company on the list: Toptal web developers group. Their famous 3% filter will give me an accurate assessment of my skills on the market. If I can pass, I will be confident enough to start a new phase of my life as a freelancer developer. Have a look on their ruby hiring guide :)

PolyConf is a two-day, single track, multi-disciplinary conference on advanced technologies for programmers interested in polyglot approach to software development. Our goal is to foster exchange between programming languages.

The conference was great, I will highlight some of my favourite talks.

Garret Smith - The Value Of Diversity And Other Lessons From Biology

We usually look to software as an engineered bridge or building (structure) but maybe we oversimplify it. Garret took a different approach and look to software as a biological organization. He pointed out that in biology, diversity is driving adoption which is the driving force of innovation. Could we change our view about software development process? Could we apply evolutionary models to our job? I highly recommend this talk, we should talk about it more.

Justyna Walkowska - Should Everybody Learn to Code?

It’s a hot topic in every community and also in education politics. Justyna is a teacher herself who taught computer science to different ages (from children to elderly). She shared her experiences and stated that teaching programming to children can be useful to strengthen their basic logical thinking.

David Nolen - Immutability: Putting The Dream Machine to Work

It was a great presentation about old and often forgotten concepts, such as persistent data structure and immutable bitmapped vector trie. He also presented some real-life examples on how we could apply these techniques to be more memory and CPU efficient. He mentioned a 50 years old paper I’ll definitely read: Man-Computer Symbiosis by J. C. R. Licklider link.

Jack Franklin - Dependency And Module Management With JavaScript

I don’t consider myself a javascript developer so I can’t really evaluate this talk but for me it was a good summary of the current situtation and the possible future. Spoiler alert: jspm.io ftw!

https://speakerdeck.com/jackfranklin/polyconf-js-modules

Alexander Solovyov - Writing Snake game in ClojureScript with React

Alexander was a genius on stage, he was live coding in ClojureScript and presented its capabilities to the audience. I liked his style a lot and he showed that how to do an entertaining and useful live coding session on a conference.

John Hughes - Testing the Hard Stuff and Staying Sane

I have already seen this presentation from John on Craft conference (recording), the concept is promising, I’m planning to give rubycheck a shot.

Irina Bednova - Using boolean logic to deal with dates

It was a short quicktalk, I liked it because she reduced a complex boolean expression with Karnaugh map. I learnt about it in high school, we used it for solving logic circuits but I’ve never thought about applying it in my everyday job :)

https://speakerdeck.com/jafrog/using-boolean-logic-to-deal-with-dates

Michał Taszycki - Commodore 64 could do that?

This was a funny talk about how we can do asynchronous programming on our beloved C64 with a neat trick. Subscribe to 64bites screencasts to learn how. Long live C64!

https://speakerdeck.com/mehowte/commodore-64-could-do-that

I don’t know if you are interested in the food on conferences but beat this:

Thanks for the organizers, everybody who came and I hope we will meet next year :)

Occasionally I still work on the OpenAM instance of Kir-Dev because they are transitioning to a custom OAuth2 provider solution and they need support for its plugins I wrote.

Few words about my use-case

My latest addition was a bigger refactor and an OAuth2 AttributeMapper implementation. With this plugin, users are able to sign in on federated sites which still use OpenAM, through the new OAuth2 provider.

With the refactor, I introduced a bug which resulted outages caused by the exhaustion of the database connection pool on Tomcat (the container we use for OpenAM). Make things worse, it wasn’t a trivial leak: it happened quite often (every couple of days) but not after some specified number of logins.

DoS it!

Okay, not exactly. I designed an experiment to generate thousands of log in operations with the help of Selenium, PhantomJS and JMeter. I learned a lot on the way - just for a hint from where I started: control Chrome windows with Selenium from a Ruby loop. I improved my setup from 15 ops/min to 50+ ops/min. I ended up with the following:

Run Selenium as standalone with PhantomJS driver

1$ export PATH=$(echo $PATH):/home/balo/apps/phantomjs-1.9.7-linux-x86_64/bin
2$ java -jar selenium-server-standalone-2.42.2.jar

Ruby script which simulates a simple login flow

 1# encoding: utf-8
 2require 'selenium-webdriver'
 3
 4OPENAM_LOGIN_URI = "https://HOSTNAME/opensso/UI/Login?service=oauthchain"
 5
 6CAPS = Selenium::WebDriver::Remote::Capabilities.phantomjs("phantomjs.cli.args" => ['--ignore-ssl-errors=true'])
 7
 8def cleanup(driver)
 9  driver.manage.delete_all_cookies
10  driver.quit
11end
12
13def login_process(driver)
14  driver.get OPENAM_LOGIN_URI
15
16  # OpenAM automatically redirects to the login form of the custom OAuth2 provider
17  user_field = driver.find_element id: "LoginForm_username"
18  user_field.send_keys "<<username>>"
19  pass_field = driver.find_element id: "LoginForm_password"
20  pass_field.send_keys "<<password>>"
21  pass_field.submit
22
23  sleep 1
24
25  # Handle the case when we have to give permission to the application
26  if driver.title.eql? "Authorize application"
27    driver.find_element(name: "authorized").click
28  end
29
30  wait = Selenium::WebDriver::Wait.new(timeout: 20)
31  begin
32    wait.until {
33      # OpenAM login url or in case of success, the protected app
34      driver.current_url.start_with? OPENAM_LOGIN_URI or driver.current_url.include? "myprotectedapp.com"
35    }
36
37    # On failure it will redirect to the OpenAM login url which displays the error
38    raise "gotcha: #{driver.title}" if driver.current_url.start_with? OPENAM_LOGIN_URI
39  ensure
40    cleanup(driver)
41  end
42end
43
44driver = Selenium::WebDriver.for :remote, url: "http://127.0.0.1:4444/wd/hub", desired_capabilities: CAPS
45login_process driver

Call the script above with JMeter OS Process Sampler a few thousand times on 4 threads and wait for failed login operations

Command: /home/balo/.rbenv/versions/2.1.2/bin/ruby
Parameter: /home/balo/code/src/github.com/vbalazs/vir-auth/do-logins.rb

About the bug

I implemented my database connection class as an AutoCloseable. My mistake was that I wanted to be smart. I made lazy connection initialization (why?!) and connection validation to make sure I handle every possible corner case in the universe. A friendly suggestion: don’t do it.

I simplified the code, from now on, it initialises the connection in the constructor and fail early if something went wrong.

Update (07 Sept 2014): @aldaris pointed out that my fix didn’t completely solve the issue. You can see his PR here. Thanks!

The second, less known solution is to use the rhc ( Red Hat Cloud) command line tool.

It has a --from-code option and you can also specify git ref however I didn’t find this in the documentation. Simply append your branch, tag or SHA after the git repository, followed by # character.

1$ rhc create-app myapp php-5.4 --from-code https://github.com/nick/myapp.git#9fbfc057

Be aware that there is a bug which makes this parameter ignored if the name of branch contains / character.

Both methods send requests to the Broker via REST API. You can read about what happens when you create an application here, thanks to Filirom1.

TL;DR: you can find my slides here:

I was a bit nervous because I have never presented at a conference before (neither at a meetup at that time) but I thought, here is my chance to take a big step forward. I kind of knew what I’d like to talk about. I got excited when I read about OpenJDK Graal and Truffle on the JRuby blog in January, I wanted to dig deeper into the topic. I sent my abstract and I began to work.

I researched two topics in parallel. On one hand, how to give a good talk and on the other hand, the main subject of my presentation itself.

How to give a good talk

I read a ton of article about this and I’m still not sure how to be a great performer. There are too much information out there, I can’t keep all the rules and “advices”. I felt I will be the worst on the whole conference… And not just on this event but every conference ever :D Then I read Avdi Grimm’s post about “How to suck at Conference talks” and finally, I was able to relax. I couldn’t thank him enough for his article. I also highly recommend speaking.io as an excellent “cut the BS” source.

On the professional side

I knew how JRuby works but this was a new topic so I started with the basics. Fortunately, I found great resources, including some good papers:

• Christian Wimmer and Chris Seaton speak at the JVM Language Summit, July 31, 2013. (link)
• Thomas Würthinger, Christian Wimmer, Andreas Wöß, Lukas Stadler, Gilles Duboscq, Christian Humer, Gregor Richards, Doug Simon, Mario Wolczko: One VM to Rule Them All. In Proceedings of Onward!, ACM Press, 2013. (link)
• Thomas Würthinger, Andreas Wöß, Lukas Stadler, Gilles Duboscq, Doug Simon, Christian Wimmer: Self-Optimizing AST Interpreters. In Proceedings of the Dynamic Languages Symposium, pages 73–82. ACM Press, 2012. (link)
• Pat Shaughnessy: Ruby Under a Microscope, ISBN-13: 978-1-59327-527-3 (link)

I had the oppurtinity to chat with Chris Seaton on IRC and on Skype. He helped me to set up a quick development environment and provided good pointers where could I research further. (Thanks!)

After that, the main problem was to how should I structure my talk. The Simonyi conference is an open and free event for everyone, organized by university students every year. I couldn’t rely on the programming knowledge of the audience, it wasn’t an exclusive event only for coders and engineers. Because the nature of the topic, I had to balance between the fans of dynamic languages and Java too. I wanted my topic to be interesting for everyone so I rewrote my whole talk like 3 times. Finally I found an approach which I could live with.

Then I faced the next problem: my 25 minutes barrier. I needed to be careful about what I will tell to the audience and what I won’t, and to stay entertaining and comprehensible at the same time. I finished my talk in 28 minutes so I definitely have room for improvement.

I feel the presentation could be better but I got some promising feedback. Hopefully I shed light on the bright future of JVM and more people will contribute to these projects.

I’d like to thank the reviews and support to Szabolcs and Tamas as well as the conference technical help to Gabor.

It’s not complicated to create tabs with custom templates in angular but I had to research a bit so maybe you won’t have to.

We will use ngSwitch directive to check the activeTab variable and ngInclude to load the right template. In earlier versions of angular it was a bit simpler to do this because we were able to use the same tag for ng-switch-when and ng-include but it’s not the case since 1.2 RC3. You can read the explanation here.

 1<ul>
 2    <li class="pure-button"
 3        ng-class="{'pure-button-active': activeTab === 'signup'}"
 4        ng-click="activeTab = 'signup'">
 5        Sign Up
 6    </li>
 7    <li class="pure-button"
 8        ng-class="{'pure-button-active': activeTab === 'login'}"
 9        ng-click="activeTab = 'login'">
10        Log in
11    </li>
12</ul>
13<div class="tab-content">
14    <div ng-switch="activeTab">
15        <div ng-switch-when="signup">
16            <div ng-include="'partials/signup.html'"></div>
17        </div>
18        <div ng-switch-when="login">
19            <div ng-include="'partials/login-box.html'"></div>
20        </div>
21    </div>
22</div>

Of course we will need a default value for activeTab so we set it in the controller.

1angular.module('frontendApp.controllers', []).
2  controller('LoginSignUpCtrl', ['$scope', function($scope) {
3      $scope.activeTab = 'login';
4}]);

Angular is pretty smart here, it loads only the default tab and if you reopen an already visited tab it won’t download the template again.

Update (04 Jan 2014): The solution above works until we don’t want to do other operations when a user switches to an other tab. I modified the original code which turned out to be more readable.

 1 <ul>
 2    <li class="pure-button"
 3        ng-class="{'pure-button-active': activeTab === 'signup'}"
 4        ng-click="switchTabTo('signup')">
 5        Sign Up
 6    </li>
 7    <li class="pure-button"
 8        ng-class="{'pure-button-active': activeTab === 'login'}"
 9        ng-click="switchTabTo('login')">
10        Log in
11    </li>
12</ul>
13<div class="tab-content">
14    <div ng-include="tabTemplates[activeTab]">
15</div>

 1angular.module('frontendApp.controllers', []).
 2  controller('LoginSignUpCtrl', ['$scope', function($scope) {
 3      $scope.activeTab = 'login';
 4      $scope.tabTemplates = {
 5        login: 'partials/login-box.html',
 6        signup: 'partials/signup.html'
 7      };
 8
 9      $scope.switchTabTo = function (tabId) {
10        $scope.activeTab = tabId;
11        /* other stuff to do */
12      };
13}]);

That means (almost¹) everything has to be pushed to the repository. We know that to commit passwords and secret keys is a bad idea. But don’t worry, use OpenShift’s environment variables! :)

The nice thing is, you can easily insert these variables into yml files, so your database config can be look like this:

 1production:
 2  adapter: mysql2
 3  database: <%= ENV['OPENSHIFT_APP_NAME'] %>
 4  host: <%= ENV['OPENSHIFT_MYSQL_DB_HOST'] %>
 5  port: <%= ENV['OPENSHIFT_MYSQL_DB_PORT'] %>
 6  username: <%= ENV['OPENSHIFT_MYSQL_DB_USERNAME'] %>
 7  password: <%= ENV['OPENSHIFT_MYSQL_DB_PASSWORD'] %>
 8  socket: <%= ENV['OPENSHIFT_MYSQL_DB_SOCKET'] %>
 9  encoding: utf8
10  pool: 5

What about custom secrets and keys?

Well, you can set custom environment variables with rhc. I made a simple text file with my variables:

1DROPBOX_APP_KEY=...
2DROPBOX_APP_SECRET=...

Then add them with set-env command:

1$ rhc set-env my/dir/openshift-env-vars -a myappname

Of course you shouldn’t commit this file to any repository.

You can test it with rhc:

1$ rhc ssh -a myappname
2$ env | grep DROPBOX

If you were connected to ssh while adding the variables, you should reconnect or check them from irb.

Tip: now you can use one repository with more than one remote.

 1@GET
 2@Produces(MediaType.APPLICATION_JSON)
 3@Path("/get/{param1}/{param2}")
 4public List<ApprovedEntrant> getEntrants(
 5            @PathParam("param1") final String param1,
 6            @PathParam("param2") final String param2) {
 7
 8     final List<ApprovedEntrant> entrants = new LinkedList<ApprovedEntrant>();
 9
10     // put elements in the list
11
12     return entrants;
13}

It seems to work as expected but sometimes we noticed weird results with the default JSON serialization of JAX-RS in Glassfish 3 (maybe Jersey?):

• in case of an empty list (!), it produces "null" string output
• if the list contains only one element, the result is the element in JSON object, instead of a JSON array with one element

Of course, I could import the official JSON Java library and do some toString() or object transform tricks, but I don’t want to introduce another dependency in my widely used domain project. I want to solve it in the webservice layer, without manual object / list JSON transformations. After some research I found the Jackson project which has a smarter JAX-RS JSON provider.

First, I added the new dependency to my webservice project:

1<dependency>
2    <groupId>com.fasterxml.jackson.jaxrs</groupId>
3    <artifactId>jackson-jaxrs-json-provider</artifactId>
4    <version>2.1.4</version>
5</dependency>

Then I had to modify my code a bit, but I think it’s quite normal and doesn’t contain ugly JSON objects and arrays.

 1private static ObjectMapper mapper = new ObjectMapper();
 2
 3@GET
 4@Produces(MediaType.APPLICATION_JSON)
 5@Path("/get/{param1}/{param2}")
 6public String getEntrants(
 7            @PathParam("param1") final String param1,
 8            @PathParam("param2") final String param2) {
 9
10     final List<ApprovedEntrant> entrants = new LinkedList<ApprovedEntrant>();
11
12     // put elements in the list
13
14     String out = "";
15     try {
16         out = mapper.writeValueAsString(entrants);
17     } catch (JsonProcessingException ex) {
18         //error handling
19     }
20
21     return out;
22}

1. Mindig ellenőrizzük a meglátogatott oldalak URL-jét. Itt is feltűnhetett volna, hogy a szokásos youtube.com helyett dl.dropbox.com van.
2. Gyanakodjunk, ha egy gyakran látogatott, ismert működésű oldal hirtelen valaminek a letöltésére, elvégzésére kér fel.
3. Csak megbízható forrásból telepítsünk bővítményeket és időnként nézzük őket végig, a már nem használtakat töröljük, tiltsuk le!
4. Ne higgyünk a kihagyhatatlan, úgy sem tudjuk x másodpercnél tovább nézni, Justin Bieber akasztása (!!!), kígyó megeszik egy embert, egyéb badarságokat kínáló bejegyzéseknek. Ha mégis nagyon érdekel, akkor legyünk gyanakvóak és használjuk a Chrome jobb gomb a linkre -> Megnyitás inkognitó ablakban funkcióját, így legalább a bejelentkezett munkamenetünkhöz nem férnek hozzá.

Néha onnan jön az ötlet, ahonnan legkevésbé várnánk. Ha van lehetőségem, akkor ebéd közben a közszolgálati csatornákat nézem a trubadúrdobozban, így történt ez ma is, az M1-en belebotlottam Szilárd Leó életéről készült filmbe.

Nagyon érdekes, ahogy a világot látta, mindenkinek bátran ajánlom a filmet. Kiragadnék belőle egy részletet: megfogalmazta a saját “Tíz parancsolatát”, ami így hangzik:

1. Ismerd föl a dolgok összefüggéseit és az emberek cselekedeteinek törvényeit, hogy mindig tudd: mit is csinálsz.
2. Tetteidet egy méltó cél vezérelje, de ne azt kérdezd szüntelenül, hogy elérhető-e ez a cél. Céljaid modellek és mintaképek legyenek, nem pedig cselekedeteid mentségei.
3. Úgy szólj az emberekhez, ahogy tenmagadhoz szólnál. Ne szavaid várható hatásával törődj, de embertársaidat se zárd ki saját világodból. Mert ha elszigetelődsz, elsiklik szemed elől az élet igazi értelme és elvesztheted a teremtés tökéletességébe vetett hitedet.
4. Ne rombold le, amit magad nem tudnál megalkotni.
5. Ne érintsd meg az ételt, hogyha nem vagy éhes.
6. Ne kívánd, amit úgysem tudsz megkapni.
7. Ne hazudj, ha nem föltétlenül szükséges.
8. Tiszteld a gyermeket. Tisztelettel figyeld szavaikat és végtelen szeretettel szólj hozzájuk.
9. Hat esztendőn át munkálkodjál. A hetedik esztendőben vonulj el magányba vagy idegenek közé, hogy barátaid véleménye ne tartson vissza attól, hogy az légy, amivé váltál.
10. Gyengéd kézzel vezesd magad az életen át és bármikor légy készen eltávozni belőle, amikor a hívó szó elhangzik.

Forrás: Fizikai Szemle; LI. évfolyam, 10. szám: 2001. október; 309. o. (innen)

Itt értesültem erről az érdekes kezdeményezésről, amelyről részletesebben az Olvasónapló blogon olvashattok.

Kíváncsiaknak egy rövid kivonat:

Mi is a csökkentsd a várólistádat?

Egy nemzetközi példára indított verseny, melynek célja, hogy az olvasásra váró könyvek kupacát csökkentsük minimum 12 kötettel, azaz annyival, ahány hónap van egy évben. Hogy legyen helye virtuális vagy valós polcán az embernek még több kötetnek, meg egy kicsit könnyítsen a lelkiismeretén.

Várólistának hívom azokat a könyveket, amelyeket beszereztél, összegyűjtöttél és már régóta arra várnak, hogy elolvasd őket, de valahogy mindig van valami fontosabb amit szeretnél olvasni.

[…]

A feladat egyszerű: Olvass el 12 könyvet 12 hónap alatt a várólistádról.

Ennek apropóján én is összeállítottam egy listát, amiket mindenképp el szeretnék olvasni mielőtt vége a világnak:

• Rejtő Jenő (P. Howard): A tizennégy karátos autó
• R. A. Salvatore: A klónok támadása
• Matthew Stover: Star Wars – A Sith-ek bosszúja
• Terry Brooks: Star Wars I. – Baljós árnyak
• Cassandra Clare: Hamuváros
• Cassandra Clare: Üvegváros
• Cassandra Clare: Csontváros
• Szergej Lukjanyenko: Világok őre
• Gerald Durrell: A véznaujjú maki meg én
• Isaac Asimov: Én, a robot
• Szabó Magda: Für Elise
• Nemere István: A fantasztikus nagynéni
• Dan Brown: Digitális erőd

Természetesen molyon is létezik kihívás.

Olvasásra fel! :)

(Ui.: Szorítsatok a vizsgaidőszakban!)

Mivel ez volt az első próbálkozásom az Inkscape-pel, ezért ne nézzétek, hogy hogyan raktam össze az alakzatokat :) Szabad felhasználni / módosítani / továbbadni engedély nélkül: letöltés (jobb gomb -> mentés másként)

Színezzétek át és nyomassátok Ti is pólóra ;)

We need your blog ID

After login, go to your Blogger Designer and you can copy the ID from your browser’s address bar.

Paste button to your blog

Add or edit a gadget in HTML mode and paste the following code:

1<img alt="Follow me on Blogger!"
2  onclick="window.open('http://www.blogger.com/follow-blog.g?blogID={BLOG_ID}',
3     'followblog',
4     'height=600, width=600, toolbar=no, menubar=no, scrollbars=no, resizable=no, location=no, directories=no, status=no');"
5  src="icons/your-blogger-icon.png"
6  style="border: 0px; cursor: pointer; width: 32px;" />

Replace the {BLOG_ID} with your ID. Of course you can use this JS snippet for almost every HTML element.

Eddig is kóstoltam már néhányszor a Snapszban, de szombaton lehetőségem volt részt venni a belga sörök fesztiválján a Mezőgazdasági Múzeumban. A rendezvény nem volt különösebben médiában, a PesitEstben és a port.hu-n láttam (Facebook event-ig már nem mentek el), de meglepetésemre, rajtunk kívül is rengetegen voltak kíváncsiak az eseményre.

Több, mint 100 fajta belga sörből lehetett választani hatot, 2500 forintért. Nem tudom, hogy a rendezők számítottak-e ekkora létszámra, mert néha nagyon szűknek éreztem a teret (pláne később az időközben túl sokat ivott emberektől). Minden korosztállyal találkozhattunk (kisbabáktól egész az idősekig), a gondos szülők belga édességekkel próbálták lekötni apróbb csemetéiket.

Az ismerőseimmel való kellemes beszélgetés közben különböző színű, sűrűségű és ízű söröket kortyolgattunk. Úgy gondolom, hogy ezt a 6 pohár sört egy átlagos férfi alkoholbírására kalkulálták, még egy pohár és nem úsztam volna meg egy kellemes zsibbadással ;) Persze azért pohár és pohár között is volt különbség (gyümölcsös vs. Trapistes Rochefort 10°), próbáltam azért az ésszerűség határán belül maradni. Sort kerítettem egy almás és egy ananászos alkotásra is, egészen egzotikus ízük volt. Az itthon kapható citromos sörök a nyomukba sem érnek.

Nyálcsorgatónak készítettem pár képet:

Szerintem ez egy tökéletes szombat délutáni program volt, megérte az árát. Mindenkinek bátran ajánlom (a gyerekeket azért legközelebb hagyjuk otthon:).

In most cases, our users would like to get a gallery with nice effects and modest manageability. This feature doesn’t come with Drupal 6, it does not support galleries well. Sure we can do a lot of magic with modules and CCKs, but it is a nightmare and anything we get as result is complicated to manage and maintain.

I have searched way too much for the solution, so I just made a decision: I will use Google Picasa Webalbums. I found a javascript library on SourceForge, that downloads the Picasa’s rss feed and creates the list of the albums and images. The result is OK, but it gives a grim sight. We can redesign it by overwriting the js’s proper sections. I have modified the code:

• it became easily configurable (with language, and config. constants)
• supports the Colorbox Drupal module
• it does not interfere with jQuery anymore

Modified pwa library (right click, save as…): download

Insert the following code into a new Drupal node:

 1<script type="text/javascript">
 2  username='yourGoogleAccountName';
 3  photosize='800';
 4  columns='4';
 5</script>
 6<script src="/sites/all/themes/yourTheme/js/pwa-no-picasa-links.js" type="text/javascript">
 7</script>
 8
 9<script type="text/javascript">
10    $("a[rel='myAlbum']").colorbox();
11</script>