In previous years, it was clear that
the Czech community was coming together, with the more experienced sharing their knowledge with the younger,
less experienced and less knowledgeable members. The focus was mainly on organisational matters, which
Vašek Stoupa (the man responsible for it all) was getting to grips with. I wouldn’t see
any problem with that. The quality of the organisation improved, the quality of the talks improved,
and the number of guests from abroad grew. Last year, several English-speaking
celebrities appeared on stage and there was a great deal of fuss about the fact that no one would understand them.
It wasn’t really that big a deal, because last
year, by my modest estimate, about 10% of the attendees borrowed translation headsets.
We had something to look forward to this year; shortly after
the end of last year’s event, it was announced that the following year’s event would be entirely in
English. We could therefore look forward to a large international turnout and people from
major companies who would bring us (“us Czechs, who are just pottering about with small projects”)
would bring us plenty of insights into how things are done abroad, how they’re done on a large scale, and
most importantly, what to avoid. As a “software developer” labelled on my
name badge, I was looking forward to it.
So there we were: the organisation from
previous years handled flawlessly, a programme with international
participation, support from Aitom covering what was perhaps a reasonable entry fee – off I went!
Three days of the programme, though I
missed the first one – the Thursday networking party. Networking there is done by a few
daredevils; the rest cluster into groups of friends and acquaintances. So I know I
can afford to skip the Thursday programme. After all, what matters to me this year is the information.
I arrived on Friday, and you know what? By the end of the day, I was very disappointed. This wasn’t
how I’d imagined it.
For the first talk, I headed to the Big
Data & IoT Hall.
V. Roček –
To be announced
The
fifteen-minute
talk was somehow cut down to about a minute. Thank goodness! With heavily Czech-accented
pronunciation, M. Bachman was introduced to us in a single sentence.
M. Bachman – (Big) Data Science
I expected to find out what it’s all about, what big data is, how it’s
handled and what’s done with it, or perhaps what the science behind it is! It took 35 minutes
to explain the graphs, which were absolute basics, and I dare say that
anyone with a secondary school education should know them, let alone in our field. And there were less than 10
minutes left for something new:
- When modelling human relationships, triads
(relationships between three people) are addressed. - Neo4j – a graph database,
which allows you to run decent queries on the graph.
I’d have liked to hear more about Neo4j, but
it was mostly just mentioned that it exists, and one query was demonstrated. Not enough. Bloody not enough.
Then I headed off with Sváča and Ondra
(let’s call them the Aitomacs) to the Design Hall, as the other talks didn’t
interest us at all. And it was a disaster. Czenglish like hell. I don’t even know who was
speaking, but he kept diving somewhere and then he must have run out of oxygen, because
he didn’t know what word to say next. “I’m sinking, I’m sinking…” will be ringing in my ears for a long time
to come. And what did I learn? Nothing! Nothing I didn’t already know.
Lunch! Well, that was a nightmare. Bland,
bland, nowhere to sit. Nothing like the other refreshments, of which
there used to be plenty.
W. Becvar – Some things you can’t
wireframe
My notes from this lecture are completely empty.
Nothing new, nothing useful.
D. Clarke – Documenting Interfaces
The first talk where I first
visited the Development Hall. And? The presenter was a disaster, more
Czech-English than Brno. Luckily, they replaced him the next day. But as for the talk itself. My expectations were elsewhere again, and the whole 45 minutes could be summed up in these
points:
- document CSS, JS, … (yes, it was about user interface documentation!),
- there is a documentation tool called KSS, primarily for CSS, LESS, etc.
(basically javadoc or phpdoc for CSS), - I (D. Clarke) made my own DSS clone, because KSS is too strict.
D. Steigerwald – Huge Web Apps
One might have expected Daniel to promote
his creation, este.js, further, but he was a pleasant surprise; his spoken
English was good and he mentioned este.js perhaps only once.
- Heavy JavaScript clients, such as applications
in Angular.js, are therefore the driving force behind large web applications. - Angular.js @steida rejected it because of its rigid
directory structure, where you cannot first divide into modules and only then
according to MVC. - Facebook React Backbone.js is the right combination.
- Use optional static typing.
- After all, you’ll always find the best mix of JS libraries and tools
for these applications at Este.
After
getting up early in the morning and feeling a bit under the weather, I’d had quite enough for the day, so I spent the rest of the time
sitting at a computer at GUG.cz and programming a better algorithm in Java than
the default one. I didn’t finish it; rather, it failed because I wasn’t able to
get the syntax right. I was programming on a computer the size of
a USB stick, where saving a file took half a minute, loading a web page took just as long,
and compiling took a minute. Not to mention that the available IDE offered almost no guidance,
let alone checked the programme before compilation. Phew.
Friday was very
unfulfilling in terms of new information, and combined with the quality of English of some of the Czech
speakers, it was quite a disaster. I wasn’t happy about it, and there was
only one thing left to do – wait and see what Saturday would bring.
Saturday began with a good breakfast (nothing like
those lunches) followed by a lacklustre talk by A. Hazdra – Service
Design in 15 minutes, nothing interesting. Just the same old things that
anyone could figure out in 15 minutes if they gave it some thought. I was really looking forward to
the trio of talks, and I finally got what I’d been waiting for.
K. Minařík, H. Král – Elasticsearch:
Beyond Simple Full-text Search
A thoroughly professional presentation
for developers. No hiccups, practical demos, live coding, brilliant speakers.
The best talk of my WebExpo this year. I knew that Elasticsearch
existed, had a REST API and was written in Java. And because it’s in Java, I hadn’t considered
actual deployment or the need to push the guys from LAMP servers into installing Java. A self-managed
virtual server didn’t appeal to me either, due to time constraints. But now,
now I’d probably like to use it, as it’s actually simpler than I thought.
- Elasticsearch is
a search engine. - It takes into account the criteria we give it (such as
post ratings; simple formulas can be written for this). - It scales absolutely without any issues, clusters
etc. - The query ‘language’ is a JSON structure, so
finding a wrapper written for e.g. PHP shouldn’t be a problem, or one could
easily write one. - For visualising logs and search data, there is
the Kibana tool. - From the large number of examples, it’s clear that this is
exactly what you’re looking for if you want to search your website (or your
data).
D. Majda – Code Reviews FTW!
A brief introduction to Code Reviews, clear,
concise, no unnecessary fluff, straight to the point.
- At SUSE, they use GitHub for code reviews.
- Pull requests are used. After a pull request is made,
someone else from the development team reviews the code and, using a checklist, assesses whether
the code is OK, or returns it to the author to discuss what
could or should be done differently. Upon acceptance of the code, the author merges the code into
the master branch themselves. - Avoid being emotional or personal when reviewing code; it is just
code. Otherwise, it can lead to discord within the team.
J. Vrána – Code Reviews with
Phabricator
Life in America really shows in the spoken
English – brilliant. A professional presentation of Phabricator and its features.
- Phabricator is the next step
in how they do things at SUSE. It organises code reviews and provides excellent
tools. - It works with the version control systems Git,
Mercurial and Subversion. - Components:
- Differential – clearly displays changes in
the code with syntax highlighting, - Diffusion – a code and commit viewer; there’s
a link for everything (for every commit), - Maniphest – bug tracker,
- Herald – notifies you of events (very cleverly)
as they happen, - and more – everything is clearly laid out on the Phabricator website.
- Differential – clearly displays changes in
Saturday was much more successful thanks to a few
speakers, though the lunch didn’t improve. On Saturday, there were no issues with English;
everyone who took to the stage spoke excellently.
I also tried out
MindBall (mind-controlled ball) in the WebExpo foyer, had a
look at several interesting desktop games for adults and drank a few cups
of very good tea from cas-na-caj.cz. I was able to have a look at (others even
tried out) Google Glass and have my photo taken in the smilebox. I saw 3D printers in action (I don’t see
anything amazing in them yet). I had a chat at one stand about extracting information from images
and video (face, figure and car detection). And that’s about it.
To sum it up, I could easily have
done without Friday’s programme. One of the organisers tried to convince me that those who aren’t
good speakers have to start somewhere and there aren’t many opportunities for that. I think there
are opportunities (conferences held at universities, start-up tours across
America, or organising a BarCamp in English). At a conference like this,
these people shouldn’t be speaking in front of others if the conference wants to
be the biggest and best in the area. The information value was much
lower for me than in previous years. I don’t know if it’s because the world has nothing left to offer, or
WebExpo was simply weaker. I would therefore
turn the first day of the conference into WebExpo for beginners! There, the large crowd of visitors aged 18–21 would probably get their money’s worth anyway. In short, when the organisation is brought to near
perfection, the content must not be forgotten.
Thank you to Martin Svačinka and Ondra Šatera for their pleasant company, and to Ondra for providing a roof over
my head. Thank you to Aitom for their financial support.
Jaroslav Moravec
22 September 2013