Is everything best out in the open?

To be open or not to be open - that is the question. Whether 'tis nobler to have software independently tested by your peers and so develop the robustness of Linux and Apache.

Or to take arms against the problems of being an acorn and rely on the market acceptance of Microsoft products with a growing range of standards.

This month's three acorns each take a different view. Procomsoft chose the open route for its software integration product, Procombox. "Open source creates the most reliable base for software," said chief executive Chi Nguyen. "Windows is not reliable because no one outside Microsoft subjects it to peer review which is independent."

Increasingly, customers will not pay huge up-front prices for unreliable software when they can pay less for reliable software from open sources, according to Nguyen.

On the other hand, Foundational Management is so impressed with C# that it believes .Net provides a better capability on which to deliver web services than any other platform, according to chief executive Freddie McMahon. Just the thing for a package to develop problem-solving solutions as web services.

Dave Overall, head of Redvers Consulting, pointed out that if you want to adopt popular standards, why not write for the standard for most of today's operational code, i.e. Cobol?. It is clearly defined, available on a wide number of platforms and can, with Redvers' XML generator, generate XML code.

All three are having to make fundamental decisions about the software on which the environment will focus. But they have something in common: they are not building a large sales organisation to take their products to market.

Redvers still depends on customers finding it. But with an 80 per cent conversion rate from trial to sales, it can rely on this passive stance for some time. Foundation wants to keep lean and is planning to use technology partners to sell on its ZDV package, as well as add components built by it.

Procomsoft laments that it did not hire a savvy sales director from day one. It is looking for a global deal to add the channel to its product.

So for all the three acorns, under the technical questions bubble the business issues: balancing product development with channel development.

CASE STUDY: PROCOMSOFT
Taking risks isn't that big a deal when you started out from a refugee boat. The IT industry can rightly be accused of being led by people whose most risky experience is driving their car over the speed limit to the next board meeting. So how about this story of the chief executive of Procomsoft for risk taking?

Born in Saigon in 1971, his mother put him alone on a refugee boat at the age of 10. The boat was robbed by Thai pirates and ran out of supplies.

The boy fainted from drinking seawater when the fresh water ran out. The boat was rescued by a passing oil tanker.

Eventually he landed in a refugee centre in the Philippines and was shipped to the US, where a family in New York state took him in. He worked hard and got into Harvard University. Those simple six sentences describe eight years in the life of Chi Nguyen.

His father was put in jail by the victorious side in the Vietnam war for collaborating with the Americans. Young Nguyen was just four. His mother knew that the only future for her child was to send him out of the country.

Setting up an acorn in the UK and tackling the complexity of software integration may seem a piece of cake after such a start in life.

Nguyen has kept moving ever since. He switched to engineering, from philosophy, at Harvard, became a programmer in a bank, was moved to London, then to venture capital operations. And finally, with two others, he decided to launch into the growing market for software and services to unite the ever growing mound of enterprise software.

The product now is Procombox. It does things such as link Siebel to Outlook, XML to Soap and MPLS, and weld email systems to applications.

Procomsoft has used just over £650,000 from a handful of private investors to build the product and start winning customers. It was too late to consider winning venture capital backing, as the market had already turned sour.

Procomsoft wants to move the target. It will package Procombox as an ASP service for medium-sized businesses that want a customer relationship management system, but do not want to host it themselves.

This will generate the cash for the third product movement Nguyen and his partners are planning: take Procombox, load it onto a Linux system, package the hardware and sell it as a commodity integration platform. This third step should be ready in two to three months.

In effect, Procomsoft is trying to do what Cisco did for the router: take specialist software, load it onto the appropriate hardware base and sell it as a commodity box optimally performing a vital task.

To get the global reach such a strategy needs, Procomsoft is in talks with IBM for a global distribution agreement. This would also avoid attacking IBM's integration software packages, such as MQ Series, head-to-head.

IBM's current commitment to Linux is in synch with Nguyen's passionate support of the open source approach.

But there is a snag: nobody, in Nguyen's opinion, has yet built a business model to make money from open source development. The Red Hat model is too simplistic, he claimed.

As a result, he is debating with his partners about what the business model and price model should be. Yew Ming Fong, Procomsoft's non-executive director with a leaning towards business, is in his mid-40s, and as focused on organic growth for Procomsoft as is Nguyen. Peter Thwaite, in his early 40s, is the executive chairman based in New York.

Between them they will have to find the right way to build the business model for open source. The future innovation of the company may not come from software development. Now that Procombox is largely completed in its present form, Procomsoft 'let go' five programmers. The innovation of the company may come more from the business model.

Whatever the model, the founders aim to grow fast. "Our goal is to play big or go home," Nguyen said. If Procomsoft is not of a certain size in two to three years, it will get stuck as a small player, he argued. Profits should flow by the end of 2003, by which time turnover should be £2.6m.

By then the company will be just under three years old, slightly younger than his other acorn - his son, Minh. And just as every tale of hardship deserves a happy ending, Minh can visit his grandparents, who are now resettled in Boston, Massachusetts.

CASE STUDY: REDVERS CONSULTING
An Overall solution for companies with a Cobol legacy wanting to generate XML.

Ted Hoff, at Intel in the early 1970s, devised a microprocessor for a project and the company realised it was the first to put a processor on a single chip. Tim Berners-Lee devised a hypertext system for colleagues in the high-energy physics community, and the world wide web was born.

Perhaps not in the same league, but in the same tradition, Redvers' Dave Overall generated an XML file from a Cobol program while working on a project, and went on to produce the first Cobol-based XML generator.

We can descend into debates about programming languages and find that Cobol is great for file handling; that over 50 per cent of working code is still in Cobol; that it is standardised; that it is rooted on mainframes and mid-range hosts; and that only a small and falling percentage of programming job adverts even mention it.

At the same time, integration is essential. It's also vital to integrate and use language tools of IT departments to produce the next round of software technologies they need.

So Overall, through his company Redvers Consulting, swung into the products business by developing the Cobol XML interface, now in version 2.3.

"There is a huge residual market out there for Cobol," he said. "Practically nobody is writing code for this market."

Cobol record structures are used to devise XML code. An additional advantage for these legacy system users is that it links to the legendary Cics transaction monitor from IBM.

So far, Redvers has sold four complete copies with a list price each of $13,500 and two parsers at $9,000 a go. This means that 60 per cent or so of Redvers' turnover now comes from product sales.

Turnover had taken something of a bashing because of the general market, and because so much had been focused on getting the product developed.

It was about £118,000 in 1999; £159,000 in 2000; fell to £93,000 in 2001; and may rise to about £150,000 in the 12 months to the end of October 2002.

The product's potential is more important than the seemingly low current turnover. It may prove to be a must for established IT operations with Cobol legacy systems which want to generate XML. Name one that doesn't.

Redvers' consultancy operations have generated the cash, often left in the company, to fund product development. Overall has adopted a conservative, even a passive, marketing approach: let the customer come to him.

The XML Interface is promoted on the Redvers website.

When one of the largest conglomerates in the world, a major UK transport company and a main US bank rely on the software of a company with three full-time employees and a handful of contractors, these customers demand assurances that the software will be there for them to change if the supplier goes under.

Redvers only offers "scrambled" source code which the user can run through their Cobol II compilers to generate the object code. No installation team is needed: it's just another Cobol application. So from the Redvers base in London it can support interest from Argentina to Australia.

If they want Cobol code readable to humans in the event of Redvers' demise, they will have to go to the escrow agreement held by the UK's NCC Group. Redvers puts the code into the NCC, which releases it to named customers if things go wrong. All of this is an insurance for the type of customer Redvers deals with.

That's all OK as long as these people find the website. But how does Overall tell the people who need his product it is there, when they are spread over the world? Such an acorn can't afford the global advertising necessary to make sure that the potential customers know of its presence.

Marketing is not the only additional concern for Overall. He's a programmer. So the extra demands of launching a product company came as something of a shock. "All those administrative things took me by surprise," he said.

To make matters simple, upgrades are free. This way every user is always on the latest version and each becomes a case study for potential customers.

Trials have a high success rate: 80 per cent of companies which have tried the product have bought it, he said. Word of mouth may yet prove the best marketing approach.

CASE STUDY: FOUNDATIONAL MANAGEMENT
It is not necessary for all businesses to be built on the same foundations. What are you told to do by everybody when launching an acorn? Write a business plan.

What did chief executive Freddie McMahon, 50, and his three co-founders of Foundational Technology do when they launched their software company? Not write a business plan.

All in their late 40s or early 50s, they have enough experience between them to know that round one of acorn building has to focus on building the product, not the business plan. They are the reverse image of the Silicon Valley start-up: great business plan, no product.

The Foundational product, ZDF, aims to solve a universal problem: a practitioner - such as an engineer, a nurse, a call centre operator or a sales person - faces a problem in their work for which they need IT support, but which is not included in the programmed processes of the systems already built.

They may have a lot of information but they do not know what to do. They will solve their problem, but the solution will die with them unless it is included in an IT system.

Rather than wait for an IT department to come up with the solution, why can't they go through a routine to describe what they want and have the software generated from a package? And the best way to deliver such solutions to other practitioners is as web services, according to Foundational.

ZDV, standing for Zero Degree Value, is a PC-based workbench that gives these practitioners a tool to build what the firm calls "know-how web services".

The user manipulates standard flow chart-style symbols to generate the solution. The solution comes out as XML and C# components. It is generated as three types of files: the application software; ASMX .Net files to describe the components; and a WSDL File to describe the components to non-.Net environments.

The three main classes of problems faced by practitioners which Foundational first focused on are: making decisions and taking action; looking up data in tables; and making calculations.

It has taken three years to build the product. The founders launched the company with their own funds in August 2001. They have used the funds from projects to take development further.

Initially, they focused on mobile applications. Talking to potential partners, they have widened their focus and enriched the ideas on which it is based.

They got on the priority lists of potential customers, but not high enough. They needed to enrich the technology to get top of the pile.

Ironically, if they had raised cash early on and plunged into the market, they would never have developed such a rich solution, according to McMahon.

ZDV is now at version 2.3 and ready for pilot applications. Novell, BEA Systems, IBM, KPMG, BT and Microsoft are in talks with Foundational as potential pilot users of ZDV, and may possibly take it to market.

The mention of .Net shows that Foundation has rather hung its hat on the Microsoft peg. "C# so impressed us, and with .Net we felt we could deliver something far better than any other platform," said McMahon.

Foundational is a virtual company. McMahon is in Essex; Michael Airey is sales director with a technology and business background based in Potters Bar; Richard Goodyear, acting as chief technology officer, is in Northampton; and Jeff Kelly, a specialist in transformation change, is in Massachusetts.

Kelly's consultancy work in the US underlined the need for the product and provided the underlying ZDV model. This work involved discovering that, although processes may be normalised inside organisations, each local group often implements the processes differently.

In other words, variety flourishes and is hard to capture. Hence, perhaps, the absence of a business plan.

See also:

Legacy Cobol lives on in .Net

Micro Focus extends software to embrace Microsoft's web services framework  29 Oct 2002

Like this story? Spread the news by clicking below:

 del.icio.us     Digg this     reddit!