9/28 session GrabTaxi

Just a brief summary on today's session.

For an app with potential to grow big, scalability is essential. The idea of micro services architecture is interesting. It builds small and autonomous units by function, letting them communicate with network calls. It fits the programming principle of "low coupling, high cohesion" perfectly well. It is a dynamic structure more friendly for addition. It also applies the decentralization concept by even decentralizing data management. The flaw comes from the external communication.

I took a special look at event collaboration. It closely implements the idea of loose coupling, so each component cares about its own business, by watching on updates needed for itself. Low coupling allows for flexibility of modification; nevertheless, it comes with complexity. Event cascade is indirect and may be difficult to spot. The communication process between components are public and incur difficulty to trace.

As for the quote from Hooi Ling "do what you really care about". Seems to fit broader than start-ups. Try to do what you care about and try to do it well!

reflections on medium

The team has introduced Medium from various perspectives and here's what I've learnt from their presentation:

1. Medium has established unique positioning in the online sharing world. It has decided to hide and weaken user, focus instead on quality content. The fact that it organizes community based on topic and content instead of relationship. differs it from other blog sites which heavily rely on interaction between users. This is ultimately important as its foundation and starting point and I feel that Medium must stick to its core value to thrive in future.


2. The group have also mentioned the technical side of Medium. Medium has put in effort to improve user writing experience. The UI is in general quite writing-friendly and it focuses on creating atmosphere deeper writing and reading. However, it has chosen to forgo some common blog functions such as categorizing comments by user and private messaging. These approaches are generally consistent with their proposition and is therefore good.

3. At last, the group have proposed possible improvements for Medium: 
In addition I have some other thoughts:
(i) Improve on the analytics aspect of user data and create value by analysing
(ii) Medium can add in ads according to specific topics

I have some add-ons on base of the presentation:  

- A natural weakness point of such products is that networking effect is vitally important. Loyalty of consumers and activeness of a tier of most valuable and influential users are key to development. Medium may seem to be even more vulnerable since there is low connectivity in the community, however, with further inspection, we will find that Medium will only attract users who are hunting for quality contents from the very start. This means it has placed itself in a niche market but in this way it's more straight-forward to offer what this small segment of users want.

 - As for the group's proposal of advertising. In my opinion, ads may be source of income, but may result in dilution of brand and loss of consumers. So this must be implemented in a careful way and under tight control. Only ads are relevant enough and consistent with its brand are allowed and in a limited space.

In conclusion, I personally feel this is a nice product targeted at a limited customer base. It is cool and brave of Medium to deprive users of many things they are so used to in order to deliver what Medium think is ultimately important. It is not a hit, it won't be one if it keeps its way of doing things, but it is something attractive and not distracted by current trends.