Last for 2017

Trying to finish that I haven't done in the last year:

1. Compiled Forked-DAAPD on Fedora 26. I had to update the autoconf in order to skip looking for libav and use ffmpeg. Also I updated the spec to include the web interface files.

2. As I have now a DAAP source I started retagging my music collection using beets. It seems that my musical taste is quite obscure as I succeeded to dag about one third of what I have collected. 

Macbook grooming

My old faithful macbook started to behave badly in the last days.
The keys O, P, 0 and - worked randomly, battery was draining rapidly and the free space was incredibly low.
To fix them I proceeded as such:

1. Low space - it seems that the Android development software was taking about 100GB, almost half of my SSD space. Just uninstalled it and, voila, lots of free space, I was able to install games again.

2. Keys - I had to open the computer, clean it reseat all the connectors on the motherboard. It was not enough though, I had also to extract the key P and move the rubber cup a little.Now it seems that the key works.

3. Battery drain - the culprit here was Chrome. Changing from Chrome to Safari increased the autonomy from about 2 hours to about four hours and this is not bad at all for a five years old computer.

I am still pondering if I should buy a new keyboard, battery and some memory upgrades. However a newer computer, something with Retina display would also be interesting.

Invaluable goodies

Today Florian asked me to find some old pictures. As I was not home I quickly ssh-ed into my server. The problem was that on text interface I was not able to browse the pictures.

I remember that I have learnt some time ago about SSHFS, a filesystem implemented over SSH.
So the next step was quite clear. Mount the folder over SSHFS.
First step was to install OSX Fuse. Piece of cake - "brew cask install osxfuse" did the job.
Then I installed SSHFS by "brew install sshfs"
Then, obviously, "sshfs nilu@myprecious:/media/Pictures ./remotefs" and voila, I have a mounted share where I am able to browse. The speed was incredibly good, felt even better than browsing in the local LAN.

November log

The Mac Minis at work were very slow, almost unusable. I managed to change the HDD into SSD in both a late 2012 model - that was kind of easy, and in a 2014 model - pretty complicated. Now they work well and can be used as build bots.

Playing with Jupyter and IPython. Mostly exploratory things.

Kind of sad. I do not feel the joy of Christmas this year. I am kind of void. I need an infusion of spirituality.

Tired. Running in too many directions at once.

How to pack bullshit

• Software Engineering Manager = timed-lead development with in-depth knowledge of one of the following technologies: Java, .NET, Ruby on Rails or PHP
• C ++ Chief Software Architect = C ++ Senior Developer
• iOS Chief Software Architect = iOS Senior Developer
• Android Chief Architect = Android Senior Developer 
• SaaS / Cloud Chief Architect = Cloud Senior Developer
• Ruby on Rails Chief Software Architect = Ruby Senior Developer
• Python Software Architect = Python Middle Developer
• Java Software Engineer.NET = probably junior Java developer
• Software EngineerQA Engineer = probably junior QA

Taken from their hiring site... So a Chief Architect is a Senior Developer. This is not false but not true either.

At least it makes it clear what they are really hunting. I think that they will easily fall under GDPR as false advertising.

Verbosus ineptias

Digi24. Bucurenci. Moartea audientelor.

Banking and Machine Learning

One of the fields where Machine Learning or any other form of artificial intelligence would bring a big plus is the banking industry. I am saying this as I feel that the current interaction I have with my bank is getting from strange to weird. Although they have the data to analyze my financial behavior, shopping preferences and cash flow, data that enables them to offer me individualized offers they still offer me all the services I do not need, which is really annoying.

In fact banks have enough data to create new services and products on the fly as they can gather enough information from a large group of people. They could create interesting options for people as overnight deposits, or create instruments with dynamic interest rate or very sophisticated investment options based on people's behavior and and risk appetite. They can learn about the human nature and offer people the possibility to use the income from the new instruments to help causes or invest in projects. I would really love to have my bank involved in some kickstarter-like projects.

By gathering contextual data there is also the option to help clients protect their investments - warn them early about market conditions, offer them products for hedging risks, offer them convenient exchange rates in case of monetary issues. It makes sense that a bank keeps the customers and their capital by offering them intelligent and customized products, not treating them in a common way.

All these require little upfront investment compared to the benefits that the banks and customers can derive from it.

I have been triggered to write this as I have been struck by the apathy of the bank clerks and of the bank itself on the current situation in Romania. They did nothing proactively to offer me any mean to hedge the inflation and the worsening of the national currency exchange rate. On the long run they benefit from the money I have on their accounts and they at least should offer a minimal hedging help. I dream having a ML financial advisor on my internet banking account that could show me some hints or offer me some customized products.

Digging in the Bytes

Did some cleaning on my NAS. Found some old tools.
I guess not available anywhere else on the internet.

Snacc for Java - IBM's Java ASN.1 compiler:
https://www.dropbox.com/sh/zl4973pthfeqmqi/AABKqbLcUysHL5eAm3P8xIqXa?dl=0

Ericsson SDS - IN JSLEE Service creation tool:
https://www.dropbox.com/sh/q812dlav214xsw1/AAD5oS6O7fwwdL3JAfoywQsBa?dl=0

Sun GCC for Sparc Servers - Modified GCC to work with Niagara processors:
https://www.dropbox.com/sh/ed58tu2ttvagj4z/AACiuZHUN7KEzripcTpTWI0Pa?dl=0

DISCLAIMER: I am not sure about the current legal status of the tools. At the moment I have obtained them they were free to download and use.

Perfect Day

Rain is gone. Sunshine through the now scattered clouds.
Autumns' green and yellow on a high blue sky.
Feeling warm. Feeling a little proud.
Listening to Lou Reed on a tube amp.
Everything is warmer.
A perfect day.
I'm gonna spend it with you.

Series

I binged this year (yes, I sinned and bought Netflix).

Stranger Things

Expanse

Ascension

Designated Survivor - fucking cheesy...

Movies

What I have seen lately:

Book of Kells - beautiful hand drawn animations and a vivid story.

And because of it I continued with:

Song of the sea - again beautiful graphics and quite a metaphoric story

Gifted - sometimes it's hard being a parent for a normal kid. (Thanks Bogdan for the tip)

Searching for Bobby Fischer - chess is as hard as mathematics:

Arrival - yeap, they still do from time to time interesting sci-fi movies.  The new Star [Wars|Treks] should burn in hell.

Java 9 surprise

I was quite shocked that JDK 9 is 64bits only.
It seems that older 32 bit runtimes are gone as most of nowadays computers are fully 64 bit.
Even Ubuntu seems to drop 32 bit support as announced in their mailing lists.

Throwing nuts

Created on 2017-09-28 11:45

Published on 2017-09-29 12:03

In Andrey Tarkovsky's movie Stalker (but also in S.T.A.L.K.E.R game and Strugatsky's book), the way of detecting gravitational anomalies in the Zone is by throwing nuts and watching how they fall down. This is how they were able to create a safe path in an area full of unknown.

Starting a new project, especially when there is not enough experience and expertise in the team or organisation, or in a new field should always start by throwing some nuts and charting a safe path. People should prepare for some failures but also for some success and use the experience gained through direct experience.

How the nuts are thrown in software projects? I reckon that prototypes that explore some of the project's goals is the answer. Small teams do the probing and draw conclusions about the results of their investigation. They would be able to rule out bad choices in technology, scope and effort. I have done this kind of nut throwing in fields like choosing a UI technology/framework, architectural styles, version control system choice and cloud infrastructure. Not all the actions were successful but at least we took some informed decisions and we avoided worse choices.

Throwing nuts also reveals what are the weak spots of the project, what possible problems might arise, and create some safety boundaries. The defined path is a better estimator of the difficulty and price of the project than any of the forecasts done when there is no concrete data on some action. It is much like "scientific method" - create hypotheses, make experiments, draw conclusions and validate the initial hypotheses.

The "nut throwing" creates proofs that can be used in steering projects. Nothing is more eloquent and informative than a proof, it is the argument that creates the "aha! moment". Quite recently I was in a situation where several architects were debating over architectural styles but the proofs made us conclude on one direction. Proofs also are an extremely important point to build upon and give to the stakeholders some insights (As a matter of fact the printing industry has a culture of proofing, soft proofing and hard proofing, that ensures that the product is inline with the customer's expectations).

Probing should be done also from time to time in order to keep the project fit and check if the surroundings changed in a way that would influence it in some way. Periodic probing of the environment also enables "plan-b" scenarios and fast recovery from uninspired choices.

Although this kind of activities seem to be an overkill and increase the cost this is not true as they make people learn by discovery and provides a way of keeping overall costs under control by avoiding the "gravity traps" that would cause extreme consequences in the project.

One year with a German car

I bought last year a Golf Mk7 SW (Variant) as the old Kalos was going out of shape and it lacked basic safety features nowadays as ABS.
Before buying the car I drove some Ford Focus, Kia C'eed, Renault Megane, Peugeot 308 and Opel Astra K on test drives. I was also considering Citroen C5 and Volvo V60 but they were kind of too expensive for me. Finally I decided to go with Golf as it had the most features for the money (it had even more features than Kia). However it was not the most pleasant car to drive Focus and 308 were far better rides (more precise steering and suspension). The engine was better than Kia's but Renault's and Peugeot's felt better.

After one year and more than 12000 km of driving the car both on city and on the countryside I can say:

1. It is not a city car - too long, hard to park, fuel consumption quite high (~9 l/100 km) due to the fact that I drive short distances.
2. Interior feels cheaper than previous versions and is prone to scratches. The interior fabric gets dirty easily and the black colour makes it visible immediately
3. Not the most comfortable chairs.
4. Trunk is big enough.
5. Lacking independent suspension on the rear wheels reduces the comfort for passengers.
6. The clutch smells burnt on moderate effort (parking manoeuvres, steeper hill roads) although it works (until now)
7. Infotainment has some bugs especially with Android Auto.
8. I expected more storage space (eg. under driver's seat)
9. Upgrades are quite expensive (eg. ACC, cameras)
10. Fuel economy on long countryside rides.
11. Good looking but not stunning.

What I am missing is the quality that makes this car unique. It's above average in every aspect but...

Latest

As I haven't written for a while:

1. Got annoyed about the stupidity of our parliament and parties. IQ test should be  mandatory for all candidates. IQ less than 90 should be forbidden. They acted as if their average IQ is well below 80.

2. Building an acrylic bending table.

3. Code Kata on http://open.kattis.com - it helps me brush some algorithmic skills.

New tricks

I have recently had to cope with mounting sun protection film on windows.
The easiest way to do it is:

1. Determine which is the foil and which is the protection. To do this use two sticky tape cuts so that you can separate the foil from the protection.

2. Clean the window thoroughly.

3. Wet the window with water by spraying it on the window.

4. Apply the film with the protection part towards you. It will stick on the window just because of superficial tension between water and foil.

5. As the protective film is in your direction you can peel it off using the sticky tape. The foil will expose its glued side.

6. Quickly spray water on the glued side of film. Use lot of water.

GET SOMEBODY TO HELP YOU HERE IN ORDER TO REDUCE RISKS OF DAMAGING THE FOIL.

7. Take the  film out of the window and turn it around. Position it carefully with the glued side on te window. As both the window pane and foil are still wet it will not adhere instantly so you can make adjustments.

8. Using a spatula (wooden, plastic) remove all the bubbles by moving from centre to sides.

9. Cut the excess foil using an Xacto knife/cutter.

 

Newest wireless keyboard

With mechanical switches.

In fact this is a CoBRA Romanian Spectrum clone inside a HC91 (another clone) case.
The idea is that it has no power supply and seems a little damaged.

Let the restoration begin :)

Back in town

Lo Tech repair

The old Bosch PSR960 battery was certainly dead. Not a big surprise as I have not used and charged it correctly all the time. I have pondered buying a new battery but the price of a battery was almost as high as for a new tool.

I had some spare NiMH batteries and a holder for them so I decided to build a new battery for the drill. Using NiMH is better than LiIon because I can reuse the same charger although LiIon would probably be more powerful and  lighter. Even with NiMH I boosted the capacity . from 1500 mAh to 2500 mAh

I removed the old NiCd cells and cut the bottom of the battery in order to make room for the holder. I have soldered the wires to the pack and placed the heat sensor on a cell. To keep them from moving (and be able to replace a cell if it would be the case I fixed them in place with some silicone.
It worked... I have a working screw driver.

Lately

- Built a 2x28V/2A PSU based on the Hiland kit I found on banggood.com. I have cut an acrylic case for it and now it looks awesome.

The design files for the case are here

- Lots of trips

- Kidney stones - excruciating pain

Missing chances

"Wasted opportunities will get back on you."

In life, as in games, one should understand a present situation and make some assumptions about future. Make a bet and try to pursue something in that direction hoping for the best while planting some "bad weather" safety guards along the way.

I haven't left Romania when I had the opportunity. I was in my comfort zone, earning well, keeping the family ties, most of my friends here. Now this is coming back on me. I feel that the country is not going in the right direction, that we think only in the present and we infest nothing in the future. And this hurts me because it will affect my daughter more than it will affect me.

On the other hand there are massive international opportunities for Romania as a country. Given the anti-globalization movement that just emerged in US there is a good chance for Romania to become a bigger player in some STEM domains.
Can we play smart this opportunity? There are some chances but overall we have to invest more. We have to invest more in Stability, schools and infrastructure.

Stability - there should be a long term plan, a bet on something that should be assumed by every political party. In this moment Romanians are playing "everybody on it's own". There is no coordination neither among individuals nor between parties, regions and such. We are kind of anarchic. The stability should be first applied in the law and taxes systems. It's kind of hard to make bets when the code of laws changes every second year and the taxation system is in perpetual oscillation. The education laws and curricula changes even more often. So we get confused students and teachers that live in a confusing system.

School - The myth of quality education in Romania was shattered for me a long time ago. Universities are close to nothingness, secondary education is in clinical death. The 4% to education budget was never in place despite all electoral promises hence teachers get underpaid so the good ones leave the system or concentrate to private lessons. The rest are petty graduates of 13th hand universities ("Spiru Haret" gang) that are not able to understand that they are teaching. They are able to reproduce the books by heart but totally unable to transmit the message or explain the content. Hence the students are again robots able to reproduce things but not creative or critical thinking.
Not being critical leads to obedience and acceptance of the status quo, one of the causes for the lack of stability.

Infrastructure - we have a fast internet that we use to download porn and dense railway on that the average speed is 30 km/h. We could use our fast networks to do something better (why there is no AWS region in Romania? Or Azure?). We could invest in our railway to make them economically viable reducing our dependence on cars and thus protecting the environment as the railways are already electrified.

I do not think that we will see any changes until the 100 years anniversary of  Greater Romania... Probably neither after. A catastrophe is needed to get us in our senses, or to finish us once for all as a nation.