How to Build an Enterprise Web App on a Budget
Ryan Carson – DropSend

Definitely the most practicaal – reall figures and good advice.

Why it’s important

What’s the big deal? you don;t have to be big anymore…this has changed the landscape

Why now – broadband; people are comformatble with web apps. They provide ‘common aps’ such as mail; hardware is cheap; open source is cheap. Does not understand why you spoend money on a .Net app – why spend the money

Whats enterprise: mass market >1000 users

On a budget??? Under £30k

Dropsend – used for sending and storing large files – without having to explain FTP. 9500 users in 2 months; 5 servers; desktops apps with API; PHP. AJAX, MySQL

The most improtant thing: ensure your idea is financially viable; use common sense; be cautious about projections; are you still in business at 65%??? Aim for profit – not aquisition; forget about the bubble – if your stuff is good it’ll work

Budget: Diff budgets for diff apps;

Branding and UI: £5000 (with friends discount!)
Development: £8500 (Plum Digital Media) – offered some equity to reduce costs
Desktops apps: £2750
XHTML/CSS: £1600
Hardware: £500
Hosting/Maintenance £800/month – for 5 boxes – maintenance as well. can afford outages. a 99.9% uptime cost is far greater. Need to know what your company needs
Legal: £2630
Accounting: £500
Misc: £1950
Trademark: £250 (do it before you finish branding – you may not get your preference)
Merchant account: £200
Payment Processor: £500

Total: £25,680
Spread out over a long time, make sure you plug into cash flow

They had a size business – the Carson Workshops

The Reality: took them about a year to save the cash; use the time to learn and work at becoming mature as a company

Building a Team on the Budget: no money, so how to build a team? Don’t go to rockstarts, go for quiet talent; offer a small percentage of equity (2-5%); ask people for recommendations; outsource – but India did not work for them.

Scaleability on a budget: buy just enough hardware to laucnh (basecamp launched on 1 server). Don;t go overboard – wait to see the success; build it so it scales; plan but don’t obsess;

How to keep it cheap: don’t spend money unless you have to; No stationery – they wasted £1000; don’t do new shiny machines; no luxuries; no frou-frou features; before you spend £25 check yourself

How to keep it cheap: make deals; give a small % away; barter services or ads; use IM/Skype; do as much as possible yourself – wireframing – wireframing-marketing-bookkeeping-copywriting; get friends to help – usability testing; shop around – the first hosting cost was £12k/month

Pessimism has its place: you’ll go 10% over budget, three months over schedule, plan and ensure in cashflow plans

lawyers are expensive! always, every single time. Terms of Service – £1000; contracts for freelancers – £800; Privacy Policy – £15; (clickbox can give some cheap legal docs); Barter! take the free consultations

Cheap software is your friend: Project Management = basecamp; Bugtracker = Trac; Meetings = Skype and AIM; Version Control = Subversion; LAMP

Cheap Hardware: cheap box for dev £200

Marketing – don’t spend money: Blogs; word of mouth; viral; writing – articles in magazines;

VC? You may need it: if you need to expand quickly; if you canpt wait to save the cash; you need a really good reason to go that way – why give up 25-50%

futureofwebapps

Shaun Inman : 10 Reasons Why You Need to Build an API

still no spelling – it may be tidied later

mint.com

What is an API: a documented means of interacting from one application to another. A successful one obscures the storage format and the retrieval process

Don’t change it – people rely on and you’ll annoy them

API = hole in the wall (ATM) – an interface with your bank
Who has an API – ‘everyone’

so why API???

1. Increase brand awareness. THe user cares what the API can be used for. Lots of people use it gives lots of brand exposure
2. Allows user to own their own data
3. Builds goodwill with developers – saves people having to do the same thing over and over
4. Perfect excuse for a community
5. Solving prgramming problems with an API in mind can improve code quality – clarify the mental model
6. Simpify the reuse of data
7. Allows others to extend the functionality of your application
8. Alternate input mechanismes – eg desktop software
9. Unanticipated applications for your data
10.Turn your programme into a platform. People build tools based on your API and it becomes a necessity.

futureofwebapps

David Heinemeier Hansson – Happy Programming with Ruby on Rails.

Still my typing sucks!

Introducing a Silver bullet…MOTIVATION

motivation has a stronger influence on productivity than any other factor (this falls into the No Shit Sherlock statements…but it’s too easily forgotten)

Motivation comes from happiness. SO you need to optimise your tools for happiness.

Beautiful code makes David happy

Your application is not a unique snowflake. most of the time you are not special. most of the work revolves around the same details as everyone else. so Ruby on Rails tries to optimise the mundane, keeping it standard – allowing the special stuff to be outside and focused on.

Convention over configuration: keep things the same where you can..

and all the note taking stops as David goes through some code….

definitely getting lost – mainly about assumptions beinbg used for conventions so you reduce what you need to write in code..and let the code framework make assumptions.

‘Flexibility is overated’…flexibility can reduce power.

‘Constraints are liberating’ – make things easier. follow people and then build on it.

Most developers find themselves with 2 voices – devil or angel. Environment influences the size of the voice. The devil says do it the easy way..you can always come back. The angel suggests you do the right thing..test, convention etc.

‘PHP is the devil’ – its easier to do the hacks and to do sloppy stuff. It’s easier to be sloppy under pressure. It bites you later.

…back to code…back to no note taking…

Why beautiful code: you feel the hurt if don’t use, you appreciate agile, you appreciate test; you can skip the vendor;

www.rubyonrails.org

futureofwebapps

Tom Coates: Designing Web 2.0-native Products for Fun and Profit

Update: The Presentation can be found on plasticbag.org – which sort of makes these notes redundant.

This was one of the most enjoyable presentations of the day. Despite Tom commenting on the URC channel through the previous 2 that all his ideas had gone, the concepts were presented in a fun and confident manner.

Live Notes: (which ‘;ve now tidied up a little)

the key characteristic of web2.0: 2 things – rounded corners and graded fills
cool font! – it’s Alega light as he states at the end (running joke of the day…who had what font)

What is the web changing into?

Web2.0: is a buzzword, a conference, a marketing challenge, a new bubble – they attempt to apply order to chaos – we lable things. We want to structure it. The web is chaos. Lots of layers in what is happening. Tech/business models/design/behaviour. We2.0 is not big enough – the church is too broad and the term will fragment.

We are moving from web pages linked to data connected by APIs It was siloed, CMS, published pages; Links, linearish….

to ‘a huge accumulation of testicles’ – the data is aggregated and joined, not the pages. data sources, services exploring and manipulating and ways that users can connect them together. to mashups – to a web of data. Eg Astronewsology – what happened to people by starsigns..and what should have happened based on their horoscopes!! (this was a mashup created by Tom of news and horoscope information) So by adding data together, they enhance each other. You get a network effect.

Every service you create can build on top of all the previous services..and the new service enhances everything that is already out there. (on the shoulders of giants….)

Consequences: massive creative possibilities, increased innovation, more competition, increased componentised and specialised services

Money to be made?????: APIs to drive people to stuff; development can be done by others – more attractive with less centralised development. Syndicated content as a platform. Turn the API into pay for services – thinks there will be an explosion in this.

If you don’t open – you’ll end up in a backwater and things will just go on around you. Open and do things faster, betterm simpler.

So what to build?

So how can I add value to the aggregate web? – Data, service to use and connection – 3 areas of focus

Data sources are in a land grab race…info and data is key to success. Owning and using data is core service.

Architectural Principles

Recommended reading – Matt Biddulph: The application of weblike design to data

Components: sources, representations, ids, distribution, interaction and rights framework

1. Look to add value to aggregate data – combine, recut, add things
2. Build for users, developers and machines. People who just want to use, who need a service; people who design and develop, how do you get them intererested; keep it consistent so machines can read
3. Start design with data and not with pages. Get data right..they feed on each other. Data right allows expansion. Design data in web like ways that can be manifested in pages
4. Identify first order objects and make them addressable. What are the core things that people will reference. like: events, people, programmes,
5. use readable, reliable and hackable URLS. be permament, 1-1 relationship with concepts; use heirarchies if ness, do no reflect the underlying technology; reflect strcuture of data, be human readable, be – or expose-identifiers (but this conflicts with previous presenters who cautioned against too much identifiers). Good URLS are beautiful and a mark of design quality! The URL love was so enthusiastic that it inspired Patrick to make this picture!
6. Correlate with external identifier schemes…or coin a new standard.
7. Build list views and batch manipulation interfaces. Three core types of pages first order (a flickr photo) slice of data to navigate btw core – photos from contacts – interface for batch manipulation – the organising pages on Flickr
8. Create parallel data services using standards All the views can be queried via APIS. eg microformats, XML, RSS, etc
9. Make your data as discoverable as possible

futureofwebapps

As reported in the Guardian, Ricky Gervais’s podcast (as hosted by the Guardian) is hitting records when it comes to downloads. Since starting in December, there have been 2.9million downloads, an average of 294,755 per episode. WOnder what the hosting/bandwidth charge is?

I’ve probably watched more TV ads today than I normally do in a week. Google and AOL both have a collection of ads that were shown in the US showing of the Superbowl last night. I’ve seen lots of build-up for what appears to be the prime placement for TV ads every year, including releases of ‘banned ads’. There’s some nice ads in that selection – I particularly like the Bud Streaker 😉

There’s nothing similar that I can think of in the UK for a whole load of ads grouped together. THere’s press about individual ads being launched (the ad break in Coronation St appears to be favourite) but nothing that can compare on such a scale.