- Provides two-ways data binding
- Second in line regarding community support
- Is a data-binding library with templating
- Provides two-ways data binding
- Has plenty of community support
- Provides testing functions
- Initial list of about 10 different tools
- We evaluated mainly the following dimensions:
A technique to isolate what to show from how to show it
logic-less vs. embedded logic
- control structures in views vs. all the logic in controllers
single-page vs. external composition
- use of partials (reusability)
- implementing the Decorator pattern to avoid having invalid HTML
before the solution, what's the problem?
- Coding DOM manipulation is slow!
- Dynamically changing HTML is cumbersome with plain old javascript
- You have to store HTML strings
- It's hard to split work among a team (Designers, developers, etc)
By Martin Briano & Rodrigo Uroz
- Uses partials -> easier collaboration
- Integrated with data-binding
- Integrated routing support
- but, it is not easy to share contexts within pages
- Documentation is not too clear
- Lots of complains form users about this
- Not intrusive into JS code. (Model properties)
- Native templating not too developed
- We improved it by implementing a custom partials managing tool
- Data binding must be explicit
- Redefines JS objects turning variables into functions.
- Changes the normal way of handling the "observable" variables
- Doesn't provide a routing module, but integrates OK with other libraries (ie: Sammy)
Use Cases:
- Modify a field within the list
- Show single student detail page
- Apply absence to a sub-set of students
- Communicate with backend using R/W REST-Services
Why a javascript framework?
- Client tiers are becoming more and more powerfull
- More computer power and memory in client software than when the web was created?
- It helps spread responsibility: the server handles the storing and business rules, the client knows how to show the data and interpretes client input with its own logic
- It balances load between server and client (servers don't get overloaded with client logic)
- Both tools are very good.
- Angular joins different modules in a consistently way.
- Knockout is easier to customize, but may lead into messy code and conflicts between libraries.
- It's true that having multiple tools makes it easier to change one small part of the app, but... does it happens frequently? Isn't it better to have a single integrated tool that is more likely to work well isolated?
- Which one to use will depend on project size and app growth expected.
- Study trending client-side libraries and frameworks
- Get close to the development experience using these tools.
- Understand pros/cons to provide support on the decision on which approach is more convenient:
- Multiple individual libraries, vs
- Share knowledge with the rest of the dev team.
- Extend use cases to do a deeper analysis on these two tools.
- Extend research to other good qualified options. i.e: Ember.js
- Discuss with the team which approach is more convenient according to projects future.