Exaud Blog
Blog
Exaud attended TECHinPORTO
Last week, members of our development team attended TECHinPORTO. If you’re curious to know a little bit more about what happened there, continue reading! Posted onby ExaudLast week, members of our development team attended TECHinPORTO – an event organized by Jumia, which is the biggest online services ecosystem in Africa, with the purpose of gathering together the best of the tech community in Oporto to freely exchange knowledge, experience, and innovative ideas. If you’re curious to know a little bit more about what happened there, continue reading!
Visual Testing – A Communication Problem
Bas Thomas Broek, an iOS
Developer at XING, delivered a presentation focused on the practice of
snapshot testing to validate certain scenarios in UI views and how
traditional unit testing replaced that approach at XING, leading to better
scalability and increasing platform independence.
Although the purpose of the presentation was actually to highlight reasons
why the snapshot testing might not be well suited for many use cases,
nevertheless we believe that it is of interest to be aware of this
methodology and relevant tools, such as
screenshot-tests-for-android
for Android and
iOSSnapshotTestCase
for iOS respectively.
Concluding, Bas tells us that there are no silver bullets when it comes to
testing practices and methodologies and, that the most important course of
action is to follow the vision that the team has built in order to define
what kind of testing results are desired, as well as, to understand clearly
how to use them as a tool to document and evaluate the progress of a given
software project.
Fast Calculations with Big Amounts of Data (on the Good Old Relational Platform)
Niko Neugebauer, a Data
Professional currently engaged as a consultant for OH22, was very pragmatic
about questioning the advantages of NoSQL and non-relational databases in
general.
In a sarcastic tone, Niko questioned the public why do non-relational
databases need to clearly state that they do the same things that any
regular database does by using the NO prefix, pointing out that it is a kind
of a lame marketing strategy in his opinion. Niko also states that the
majority of NoSQL databases rely on SQL primarily, pushing it a bit further
by suggesting that if these databases so like to employ the “NoSQL”
designation, then we should have in the opposite side of the scale “YesSQL”
databases as well.
The presentation went on by showing that a traditional relational database
could be as fast as a non-relational database in performing big data related
operations using the Batch Execution Mode for queries. That point was
convincingly concluded by running a few live queries utilizing BEM for
tables with billions of rows.
Architecting for Machine Learning
Julio Faerman, one of the professionals responsible for Developer Relations at Amazon
Web Services did a quick walkthrough of the Amazon Machine Learning
capabilities and APIs.
He started by showcasing the
Amazon Rekognition that
is well capable of analyzing and successfully identifying objects, and
people present on a photo or video. He did a couple of tests with photos
taken on the spot and, it was clear that in low light conditions there was
still room for improvement, nevertheless the technology looks
promising.
Another interesting item presented was Amazon Polly, a text-to-speech tool
that caught our attention when Julio shared that there are some radio
stations in Japan using it to replace broadcast presenters, as it has got a
lot of well-engineered features to actually generate a human-like speech.
Then he proceeded with some funny examples, such as emulating
whispering.
Amazon Lex was also shown, an interesting SDK that provides developers with
the same set of speech interfaces that are currently used by Amazon smart
assistant, Alexa.
The presentation also featured an overview of the Amazon Comprehend, an
evolving tool to analyze the meaning of any given text, for example
determining if the overall tone is positive or negative.
Finally, Julio went on to the Amazon Transcribe, an automatic speech
recognition technology that can turn speech into text in any of the
currently supported languages.
Most importantly, all of these fascinating services have API ready to be
used by developers for various apps without any constraints imposed by
Amazon, thus making the Machine Learning capabilities easily available to
everyone.
Blockchain is not what you think
That was one of the presentations that concluded the TECHinPORTO
conference, where Bernardo Vieira, a QA automation tester at Altice Labs,
gave a general explanation about what blockchain is and how it works,
including some of the popular use cases, as well as associated
myths
and misconceptions. Providing a bit of history regarding the origins of this
technology, Bernardo also emphasized the fact that blockchain does not apply
only to cryptocurrency, as many people might think.
In fact, there have been several important innovations and technological
improvements due to the increasing blockchain adoption, which was indeed
fuelled by the widespread success and popularity of Bitcoin and other
cryptocurrencies. He also highlighted examples of some of its current and
potential applications, such as creation of decentralized applications using
platforms such as Ethereum, and forming
the backbone of a new protocol for the Distributed Worldwide Web called
IPFS.
Conclusion
In the end, we found the event very interesting and enjoyable.
Although Exaud is focused on mobile and embedded development and
TECHinPORTO is more cloud-oriented, we ended up enjoying multiple topics
relevant to our expertise such as Machine Learning, Non-relational
Databases, Blockchain and Testing.
It is always a pleasure to broaden our knowledge and learn more about new
fascinating technologies and growing trends.
We will certainly make another appearance in
TECHinPORTO in the upcoming
2019!