NERDS> Millennium Edition - Progress Report
This report is brief, and essentially summarises an internal report, and many Minutes of Meetings - without divulging design details.. I must remind readers of this report that many aspects of the new design remain secret or confidential, an explanation of some of the prosed features can be found here.
The purpose of this report is to outline our current progress, call for assistance with testing, and ask for outside help with some ‘development' hitches.
A point: When you're building one of the largest communication centres on the net, you're best off keeping it secret until your final release, because everybody wants to help otherwise, including: FBI, CIA, MI6, hackers and probably the mafia.. This is when you go absolutely paranoid about security, first thing out is the MessySoft server.. next you design a system of moderation, and sophisticated collection of ‘intelligent' bots who watch the network, and network architecture that handles mail booms and demand escalations on the fly..
What we're saying is, stop the moaning please, this design is tricky stuff. Consider a mailing list with 400 posts per day (it has too many members) it just doesn't work - because no list member would ever read them all. We are having large problems with economies of scale, and member distribution. If you have 4 mailing lists rather than 1, this would mean 100 posts per day? Nope, doesn't work either, 2/3 subscribe to all 4 of the mailing lists...
The above paragraph explains why we have released details, and raised interest about the new version, against the rule of secrecy. There is much critical testing that must be undertaken, to expect real results we need to run the tests with the scale the final network will endure, and for that we needed your help. Don't forget the problem with the existing network, the reason for the subscriber limit, is under capacity, or as we like to call it ‘member distribution'.
The design for NERDS> II (NERDS> Millennium Edition) started in October 1999, we thought we had a large task ahead of us, but we just didn't realise the extent to which ‘large' really meant. The existing network isn't really one network at all, it is many systems patched together. We wanted to build a common, expandable and distributed system that linked the different types of services together via a single point. We wanted to be able to ‘flick a switch' to introduce a new channel, thus channels or rather sub-channels could be created on the fly by the system users, and provide a secure hieratical user base, over a continually expanding and distributed network.
Along with the design implementation, we also defined many new services that we wished to incorporate into the network. We realised that in the future we would need to introduce other new services, and even change the way current services work (or remove them completely). For all we know two years down the line Internet users will consider IRC obsolete, using some form of distributed video link?? Everything had to be flexible, on a scale in which ‘flexible' is not normally recognised.
After two months of design (well, one month of specification, and one month of design) we realised that we couldn't design all the components on our own, and that testing and trialing is a necessary part of our design process, throughout the design - we can't afford to get the design wrong, there are too many aspirations, and potentially the new network could be worse than the old one if not effectively tested and proved.
After some deliberation I created the new site you see today, we contacted organisations and friends who had experience with the technologies we were working with. The project team grew, and grew, and grew.. Over the Christmas period whilst everybody had time off an astonishing amount of work was achieved, the core design, the main databases and interface components were completed to a stage where a running prototype was possible.
The new century arrived, and for some unknown reason (at the time) this new front portal site was extremely busy.. New surfers we thought (pick a popular keyword for searching the Internet 'nerds'?), or members returning to their Internet connections at work.. Well we apologise, seems that everyone thought that ‘NERDS> Millennium Edition' means that we would open up the new site on the darn of the next Century.. pah! We had barely begun, we had designed a distributed database, and a proposal interface...
The design team, nine members in all now, hasn't changed much since Christmas, although the pace has slowed down. Once the centre was complete, we split up to work on individual components, adding them to the central design, and reviewing progress twice every week. Considering our team, with the exception of audio & video technologies, there doesn't appear to be any skills gap, we have a highly skilled and experienced team - although, they remain anonymous because some of their employers wouldn't like it - say no more!
We have just evaluated our entire design process, and planned a schedule for development. We have decided, whether wise, or not, to design all our own services, from IRC, BB, e-mail service, file store, search facilities, etc.. Why, because it allows us to incorporate the services together, and modify the design of each component ourselves. This allows us to provide better search facilities, provide a single system of moderation, more effectively control the distribution of load. Thus provide the flexibility, security and load requirements we are looking for.
So what is the schedule? - well that is classified, and very difficult, if not impossible to create over any substantial time period. Our design team is made up of volunteers, quality control and security are of the utmost importance, and we would like to be able to implement more than one final network supporting different languages and countries. What is more, although we have specified, and keep extending our specification, we are unsure what is involved to design some of the components, or at least we haven't thought about them yet. Not forgetting that the network we are designing them for (The Internet) is rapidly changing, through design(protocols & standards), and the way it is approached and used.
The biggest task to undertake is implementing the audio and video technologies, the problem is that there isn't a complete working model to work with. DVD, G3 and other such digital consumer systems are only just being introduced, and there still isn't a settled global specification. There are also very few applications build for the web, and none of them support multiple users very well, and they all use proprietary standards. To implement solutions effectively we will have to design the server-side systems, standards and some protocols for ourselves, this also means designing the clients too - this you may know has already begun.
Without divulging too many details, I hope the above excites your mind. We have made a great effort so far, and much has been achieved, even though we have little to show for it. If you think you can help, if you have designed some of the components that we are mostly likely going to have to implement (very cleaver I know ;-), please don't hesitate to contact us. NERDStart> 2000 Team
BTW: You'll find further hints about this project within my previous two editorials.