As a software manager, I lived with the wrath of Gantt charts and Microsoft Project schedules for too many years. They get in the way of making progress, and they get in the way of your relationships with your staff, peers, managers, executives and directors, and most importantly, your customers. The world doesn’t need another scheduling tool. Have they not been listening to Kathy Sierra?

Setting up Junctionbox Media helped throw Gantt out of my life. Furthermore, the tools and techniques we use now – major shout-out to Ruby on Rails – almost demand that you ignore Gantt, because there are easier ways to schedule and prioritise work.

It’s somewhat ironic then, that I discovered a thread today discussing how to implement Painless Software Schedules (PSS). This article by Joel Spolsky was written back in March 2000, and it’s still relevant today. It suggests using Microsoft Excel as a basis of scheduling, and includes some useful advice on what not to forget and how to estimate. Unfortunately, it’s unlikely that a job that needs to be scheduled is wholly understood so as to make such a schedule valuable. Hence the rise in agile software techniques, which we use (and it’s feasible to use PSS for each timebox).

But where’s the irony, you ask? It’s here: discussing how to subtract two columns of numbers in Excel.

I read today that CouchSurfing has closed. CouchSurfing was a site that allowed people to offer their homes for free accommodation to travellers. It was three years old and had some 90,000 members.

It closed due to catastrophic failures in some of the operational aspects of the site:

• a hard drive crash that was avoidable
• incorrectly executed incremental backups

I used to work for a company that didn’t have proper backup procedures – not only for itself, but also for its major product. In both cases, checks weren’t performed to ensure that it was possible to recover from backup.

Backups are rarely seen as a necessity in small businesses and few organisations are committed to the time and effort necessary to ensure that its complete and correct. It’s essential and critical. Because, as the CouchSurfing experience demonstrates, if something goes wrong it can close your business.

We’re unlike most businesses. We have an implicit advantage in that we develop locally and deploy to remote servers. In a sense that’s an automatic backup, although that doesn’t give us the ability to rollback. And we rely on the owners of those servers to back them up in accordance with their terms of service. Service availability and backups are less than ideal, but commensurate with the expectations of our customers. Even so, the more we use other companies, the more we are coming to realise that one day we will need to do it all ourselves.

We use Subversion to commit and manage changes to our work, on our local servers. Thus, our servers have the entire lifetime of each project, with individual employee’s computers retaining just the local ongoing work.

But that’s not all. We also backup our servers. A custom-developed automated process uses the internet not only to ensure that we have both an incremental backup, but that we also have an off-site backup. This gives us, in total, three or four distinct places that our work is stored.

Yes, it’s been tested, and yes, it’s saved our business.

Much of the advice given about running an online store discusses issues related to providing attractive, secure, reliable and trusted online presence. It doesn’t include advice about selling. That’s because it is assumed that people selling products online already know about selling offline, using traditional outlets. Because building a website does not automatically build a business.

When setting up a physical store one has to:

• determine if a market exists for your products or services
• tell your market that you exist
• get people to choose you over your competitors
• set the prices
• determine the shipping methods and prices
• staff the store
• define the policies to you apply to the store, staff and customers
• present and describe your products
• achieve a sale in the easiest way possible
• encourage related and repeat sales

Each consideration in the above list is about persuading people to buy. There’s no other reason for a store to exist, except to sell stuff.

Building an online store only provides the basic mechanism for a store. A website does not find customers and does not force people to buy its products or services. It cannot be built and forgotten about, whilst you wait for the orders to magically arrive. It requires work and commitment.

This from an article on A List Apart:

Every business that settles for a poorly built badly managed website is doing nothing more than pulling the web down with them.

The future of the web lies with the web developers and designers; but what happens when the jobs go to the developers and designers who don’t care about the future of the web? The web remains stagnant or, worse, travels downward.

Sounds much like something I wrote recently.

Trouble is, the article, entitled Cheaper Over Better: Why Web Clients Settle for Less, was from 15 July 2001.