Mabble Rabble: January 2020

Office politics, blame games, misrepresented costing, and misguidance are the result of why organizations still continue to use Azure. Microsoft has never been great with networking or anything that relates to internet services. The below highlights the many reasons why one should stay well clear of Azure.

Cluttered, unclear and inconsistent documentation
Active Directory use is everywhere so a nightmare to manage with resource groups
Tight coupling is literally in every aspect from implementation demos to service integration
Riddled with bugs across the service stack
Sometimes even Azure Consultants get the horrible blue screen while doing their demo on their shiny Windows machine
Nothing ever works properly, lots of unnecessary time wastage interfacing with their UI
Even the UI for their services violates basic human factors of interaction design
Metrics are totally useless, unresponsive, and meaningless
Support is useless, one could be spending time being re-routed and re-routed
Lack of consistency across services so the entire platform has bad integration and lacks coherence
Constant pestering from sales reps to convert to their enterprise plan without applying any said discounts
Platform uses ancient desktop methods which don't quite work in the Cloud
For every hosted service they provide, one can find an even better open source solution
All their implementation demos have a bias towards Windows, C# and Powershell use cases
Their AI and Bing services are utterly horrendous, even Azure Consultants think Bing sucks
In fact, one has a slower time to market while being less productive and efficient
All their services have a Microsoft bias over Linux use cases
Usage costs on services and breakdowns is rarely very clear and transparent
Microsoft has a way of overly complicating simple things by creating UI experiences that don't work towards usability and responsiveness
Often people find moving away from Azure to AWS a pleasant experience
Azure is like an evil maze controlled by a circus clown
Too many services with little quality assurance that rarely work properly
Almost every service experience seems like a half-hearted effort
Increase in support costs as a result of an unstable and unreliable platform
For many cases 'this only works on Windows' continues to apply on Azure
Most of Azure's Quick Start Guides take longer to understand, will be easier to Google for even better guides outside of Microsoft realm, which not only will be quick but also clear
Almost every implementation demo seems to focus on Visual Studio, what about other IDEs like Intellij, and others.
Sometimes services may not shutdown when Azure says it has done it
Deployments and Builds are not simple but deceptively cryptic
Lots and lots of hidden fees and the costs pile up
Missing documentations that will make one pull out their hairs in frustration
Things break, aimlessly, and constantly like the entire platform is running in beta mode
Many of the Azure Services are box standard Microsoft desktop software that has been re-purposed and re-mapped for the Cloud

22 January 2020

SemEval

Hire a Robot

In the world of work, job specs are changing face. When you see job spec requirements that say:

"linear execution and lateral problem solving skills"

You know you need a robot not a human.

Do most people even know what linear vs lateral thinking is? In fact, there is even a third type of thinking called diagonal. Keep job descriptions to the point, honest, simple to read, and simple to understand. And, no need to embellish on the role - candidates have a way of finding out the truth.

21 January 2020

Open Data Platforms

DSpace
Eprints
Invenio
Zenodo
CKAN
DataVerse
DKAN
Junar
OpenDataSoft
Semantic Media Wiki
Swirrl
Socrata

Pachyderm

DeepWalk

Network Repository

Deck.gl

20 January 2020

Apache Superset

18 January 2020

Linguistic and Web Annotations

NLP2RDF
Distributed Text Services
LAPPS
CoNLL-RDF
Text Encoding Initiative with LOD Linking
Interlinear Glossed Text
LD4LT
NIF
OntoLex
Web Annotations

14 January 2020

Ontology Design Patterns

UCLA Semantic Web Library

12 January 2020

KNative

11 January 2020

Linked Open Datasets

Linked Open Datasets
LOD Cloud
DataSet RDF Dumps
Linked Open Vocabularies

RDFHDT

9 January 2020

Knowledge Graph Embeddings

Knowledge Graph Embeddings
KG Embedding Survey
KG Embedding Survey 2
Comprehensive Survey
KG Survey 3
KRL Papers

DeepPath

Fonduer

Knowledge Graph Frameworks

GNoss
WhyIs

NanoPub

Depot

GroundAI

Model Zoo

6 January 2020

Lore

3 January 2020

Grafana, Prometheus, and Netdata

2 January 2020

Google Cloud Is It Really GDPR Compliant

Google has notoriously been a consumer and a provider of data through its proprietary search technology. In fact, they have even violated EU laws on many occasions and are still under scrutiny for outstanding cases. Also, they have even been under scrutiny for violations of privacy laws with patient health records at the DeepMind-NHS collaboration. Why then should we assume that Google Cloud is GDPR compliant when they cannot keep their own storage and processing of data in check? Who is to say that anything that Google Cloud Customers store and process would not necessarily get leaked into other parts of Google Services like Search. Just to make a 'right to be forgotten' request or to request an update to an outdated data on Google can take an eternity in some cases. In fact, image search is one such service that has outdated cached index of links of images which don't even exist for many people at the source anymore. But, who is to say that maintaining outdated image links is not really an issue but a feature to assist in their third-party initiatives in building computer vision datasets and models. One has to be weary of cloud providers who claim to be GDPR compliant while violating such mandates across their own systems of accurately storing and processing data within the bounds of consent.

CNIL

Digiday

CCPA

Third-Party Violations

DeepMind-NHS

Five NHS Trusts
Google Fined for Privacy
List of Data Breaches
Google Litigation
European Union vs Google
US International Trade Commission