See also Commercial Portals | Cookie-Cutter Portals
You can see from the hard evidence below, that using open source tools to develop custom Web portals is much more viable than using commercial tools to maintain Windows open source portals even though Windows open source portals are free to download. The crunch is the ongoing support for commercial tools is much higher and less lucrative than open source tools due to open source tools (not application software) having a much LOWER overhead and much better support infrastructure!
Microsoft and Apple Mac based servers or languages or cloning custom sites are not viable options for small developers like ourselves. They require expensive support by trained Microsoft or Apple engineers which our customer base will not pay for so we are lumbered with the huge costs of support for little guys trying to enter big guy territory - it does not compute! The same happens with CodeCharge which produces very complex obtuse designs which are hard to maintain unless costs blowout or deadlines are blown.
We prefer original custom work on Linux on x86 any day! We are not a middle sized application company or a hosting company with technical support laid on. We are developers who build economic, humble systems that WORK! We have perfected a project management system that controls people who try to leverage us for something we are not capable of doing due to our small size! We will exit if the project gets too hard for us and handover to another developer to complete the work.
Our project management system documents all designs by linking modules and data dictionary entries and generating documentation which did not exist. This has proven very useful for handling other developers or customers and helps us to control a project from our end despite the customer not doing the documentation. We just do documentation on our own tools and setup our own offsite testbed so we are independent of sysops. It is a great freedom not to have to rely on another manager for vital information and stop delays.
On a deeper note, we avoid poor project management and the pain it inflicts on us. If the system is designed on the wrong premise (e.g. clean data), then redesign will be costly but effective. Network support and poor quality infrastructure (CPU, RAM, hard drive and bandwidth) can drown the best programming skills.
Basically, we engineer our products ourselves to save the cost blowouts of modifying and created by using other people's software and hardware infrastructure which looks cheaper but in the end is quite costly unless you wrote and set it up yourself. Hence we have our own test beds and don't rely on other people's infrastructure.
The trend the past 2-3 years (2005-2008) is for people to use open-source portal software and get a developer like us to modify it or set it up. Needless to say, this has eroded considerably how much we can make out of developing or installing a portal. As ever, the customer drives down the price as products become commoditized. This has driven us into developing mobile and mapping applications.
| Market | Status | Sample | Application Software | Platform Technology | Application Licensing | Platform Licensing | Costs | Overheads |
|---|---|---|---|---|---|---|---|---|
| Commercial Portals | Dead | Dating Service | DateMill | PHP, MySQL, Linux, x86 | Commercial | Open Source | 90% support, 10% design | Hard to maintain or configure, complex software design, poor documentation, good tools, good support forums |
| Cookie-Cutter Portals | Revived | Community Group | DotNetNuke | .NET Framework, SQL Server, Microsoft, x86 | Open Source | Commercial | 80% support, 20% design | Brittle configuration, buggy software, poor software design, constant upgrading, costly or poor quality tools, messy time consuming support forums, domain transfer support |
| Commercial Portals | Expired | Unnamed Business | original custom | PHP, MySQL, Linux, x86 | Commercial | Open Source | 5% support, 80% programming, 15% design | Solid, few bugs and bugs quickly fixed, quality software design, few upgrades, cheap quality tools, quality support forums |
| Commercial Portals | Dead | Unnamed Company | clone custom | PHP, MySQL, Mac OS X, G5 PowerMac, Airport | Commercial | Commercial | 90% support, 5% programming, 5% design | Poorly documented tools, poor security on networks, hard to backup and restore, difficult to port tools to, expensive, hardware unsupported in near future, obtuse support forums |
| Commercial Portals | Dying | Unnamed Company | generated | PHP, Postgresql, Linux, x86, CodeCharge, LAN, Remote Desktop | Commercial | Commercial | 90% support, 5% programming, 5% design | Poorly documented tools and design of site making it very hard to fix, hard to manage using CVS due to obscure undocumented builds, difficult to access databases due to badly configured servers, difficult security restrictions on servers stopping testing, secretive unhelpful untrained fellow developers and network manager, expensive and hard to move IDE licence making it difficult to setup tools, obtuse support forums |
| Commercial Portals | Dead | Unnamed Company | Custom import of 1 million records in ASCII format merged with SOAP feed | PHP, MySQL, awk, C, Python, Linux, x86, ssh, SOAP | Commercial | Open Source | 15% support, 80% programming, 5% design | Change in specifications due to dirty data requiring a rewrite to clean data before import which doubled costs but shortened development time; a huge amount of sysop on MySQL and speed of CPU, size and speed of hard disk and RAM and network connectivity due to old hardware and erratic networks; optimising PHP and C to speed up import from weeks to hours and stop memory crashes and disk filling up with logs |
| Commercial Portals | Unfinished | Unnamed Company | Translations contact and project database on intranet | PHP, MySQL, Linux, x86, Javascript | Commercial, Open Source | Open Source | 10% support, 80% programming, 10% design | Complex data structure and old data led to change of menu design using Javascript and problems importing data due to poor format of data and low technical and information management skills of user; tech support drowning out time for development and killing job. |
NB: Any sites without a link have been redeveloped by another developer or gone bust.
Created: 25 Jan 2006 13:35
Last Updated: 12 Oct 2008 14:30