Some quick thoughts about the comparison between old Enterprise projects and the new style of web projects:
Category | Enterprise | Web |
---|---|---|
Project style | Waterfall - Design, Develop, Test, Deploy, RUP | Agile - rapid iterations |
Innovation | Top-down (Business Requirements) | Bottom-up (Google 20% time, Communities) |
Team structure | Vertical - developers code complete spikes from web to database | Horizontal - front-end team is split from (and a client of) back-end team |
Software | COTS Products - expensive, well-supported commercial products (ATG, IBM, MSFT) | OSS Frameworks & Patterns, Community-led initiatives, Bespoke 'glue', products for specific functionality |
Development Environment | Restricted software tools, homogenous environment, complex approval process | Heterogeneous, open, embracing new technologies (best tool for the job) |
Development Languages | Java, .NET, C++ | PHP, Perl, RoR, Python, JavaScript, … |
Buzzwords | SOA, SaaS, Compliance | jQuery, Memcache, LAMP, JSON, API, OAuth, Mashup |
Web Services | SOAP,WS-* | REST, JSON, POX |
Documentation | Offline - licensed developer accounts; formal training courses | Online - updated on the fly, available to everyone |
Architecture | N-Tier | Distributed |
Infrastructure | Best-of-breed hardware, design against failure | Low-end hardware, expect failure and design accordingly |
Database | Normalised, strict schema, referential integrity | Denormalised, dirty data, lazy-loaded |
Data access | Direct – ODBC, JDBC, ADO.NET | ORM Framework, Cache, Services |
Requirements | Secure, Scalable | Flexible, Fast |
Hero clients / employers | Government, Banks, Corporates | Google, Yahoo!, Amazon, eBay, Facebook, BBC, Twitter, Next Big Thing |
Success Criteria | How big can we grow, how secure is our data? | How fast can we react to change, how far can we scale? |
No comments:
Post a Comment