At Cyberport, home-grown start-ups attracting attention include GoGoVan, which links freelance truck drivers with people who need goods delivered, and Coachbase, a developer of digital tools for sports teams. Then there's AfterShip, which operates a global package tracking system for online retailers.

Yet a weekday visit to Cyberport in Pok Fu Lam still has a surreal feel. It is a thoroughly modern complex, filled with gleaming glass and LCD screens, but there are few people in its offices and public areas. To critics (and there have been many from the moment the government awarded development of its prime waterfront site to PCCW without tender), such scenes only reinforce views that Cyberport was a misguided bid to springboard the city to the forefront of information technology in Asia by building a facility for related businesses.

IT sector legislator Charles Peter Mok, who was one of the advising consultants, says the project was hampered by its bad reputation, describing it as an "illegitimate child".

"That stigma had to do with the way the project development was given to PCCW without tender. But that doesn't mean that Cyberport was not doing its job of being a project to support the development of the IT industry. Quite the contrary," he says.

Whether inventive engineers or software developers could have achieved as much without the HK$13 billion investment in creating Cyberport is another matter.

"It's easy to see that good things have come from Cyberport - there's no doubt about that," says entrepreneur Joshua Steimle, a co-director of the local chapter of tech community Startup Grind. "The question is whether those good things have been worth the cost."

The project has had a modest impact on the start-up scene, he says.

"But if Cyberport wasn't there, these start-ups would find office space somewhere else in one of the many co-work spaces that are popping up every week. Is it helping? Probably. Is it worth it? Hard to say."

Although global giants such as Microsoft and Cisco came in as tenants, the facility hasn't exactly been a hive of activity. Offices were mostly empty in the early years, as many firms found the location inconvenient; until 2008, even data centre occupancy hovered at 20 per cent.

Herman Lam Heung-yeung, who was appointed Cyberport chief executive in 2010, feels criticisms about the project's modest impact are unfair as the facility opened when Hong Kong was in the midst of recession in the aftermath of the Sars outbreak and the 9/11 attacks.

These days, 90 per cent of offices are leased out; over the years its creative microfund scheme has awarded seed money (HK$100,000) to 108 projects. By March 2014, Cyberport management had accepted 210 start-ups into its incubation programme, which offers among other things, technical support, subsidies for professional services, training and staff, as well as 24 months of rent-free space if they locate in Pok Fu Lam.

About 87 per cent of companies are still operating 18 months after "graduating" from the incubation programme, but the survival rate drops after three years to 58 per cent.

"We understand that Cyberport is too small," says Lam. "Hong Kong is too small."

His solution was to launch a HK$100 million three-year plan focused on initiatives that are not dependent on Cyberport's physical facilities. "We need to expand 'virtual' Cyberport."

It's telling that Lam is focusing on virtual Cyberport's mix of training, internship and incubation efforts along with the microfunding scheme. Promising start-ups such as Aftership have preferred not to move to Pok Fu Lam, even though they have benefited from its microfund scheme - "We used that money to build our first prototype and secure our first angel investor," says co-founder Andrew Chan - and went on to sign up for its incubation programme.

While Cyberport gets most public attention, Science Park has developed into a campus covering more than 220,000 square metres, with offices and facilities that now bring in a working population of more than 10,000 people.

Part of a three-pronged scheme, Science Park was conceived to work in concert with the InnoCentre, an incubator based in Kowloon Tong for up-and-coming design companies, and industrial estates in Tai Po, Tseung Kwan O and Yuen Long. Science Park would handle research and development, while the InnoCentre design and the industrial estates would provide support and manufacturing.

The architect of the plan, former University of California Berkeley chancellor Tien Chang-lin, envisioned the scheme to galvanise growth in sectors from green technology and electronics to telecommunications and precision engineering. Although it's difficult for locally based manufacturing to compete with the mainland, land leases from the industrial estate have provided steady income for the Hong Kong Science & Technology Parks Corporation, which manages all three arms.

The funds have helped finance expansion at Science Park, which is now in its third phase of construction. There is a buzz about the place, and lunchtime brings crowds into its cluster of restaurants.

Still, the Science Park's greatest impact may stem from its "software", that is, its incubation programmes.

Start-ups that are accepted are given sizeable rental subsidies, access to park facilities, and help from business mentors for periods of between 18 months and four years. The park also provides a range of support services from help with recruiting talent and dealing with media to access to angel investors.

Of the 309 ventures to graduate from its programme so far, 233 are still in business. Another 145 companies are in the programme.

Sengital is a typical example. Its chief executive, Alan Lam, started the business based on motion-sensor technology he developed while studying for a PhD at Chinese University. A Science Park executive encouraged Lam to apply for its incubator programme after he won a couple of start-up contests.

"In the beginning, I didn't quite understand the kind of benefits, but I thought that this was a nice place and it was near my dorm," he says. Although an investor offered to help with manufacturing on the mainland, Lam says he didn't have a network with customers.

"My mission was to find out how to handle a complex organisation and find out how to sell the product," he says.

Through Science Park, he met successful CEOs and enrolled in basic business classes, hooked up with distributors, and learned about making his product commercially viable.

Lam wound up designing a gaming controller for PCs, but could not make enough from the first orders to pay his employees. So he went to the park management, which gave him 800 square feet of free office space and a subsidised rate for extra space. Lam received help from the park management that sometimes verged on hand holding, for example, when hiring new staff.

"We just listed the kind of people we wanted and the salary range and they passed it to the university people, collected application letters, filtered them and sent them back to us. They even gave us a conference room for the interviews so the students would come in and think this was a good company."

Sengital repeatedly fell on hard times, and was saved repeatedly by Science Park management.

Today Sengital employs more than 30 people, and Lam has even started to invest in other companies, among them a restaurant in Science Park.

All the same, this experience begs the question: are projects like Science Park keeping alive start-ups that would not succeed in a more competitive environment?

Many start-ups have chosen to operate outside the umbrellas of Science Park and Cyberport.

Ambi Labs, for instance, applies machine learning and big-data techniques to making smart household devices (first in the pipeline is an air-conditioning system that can be controlled by your smartphone). Its CEO Julian Lee Shang-hsin says they considered joining one of the quasi tech complexes but ultimately decided against it.

"For a start-up, all resources are constrained. Our biggest constraint is time, and given their locations and travelling times we couldn't justify it."

Former legislator Sin Chung-kai, who represented the IT sector, is blunt. "Both projects weren't failures, but I wouldn't say they were huge successes. Hong Kong's economy has not been able to diversify. It has become even more reliant on the four major sectors of financial services, trade services, professional services and tourism," he says. "We have been unable to turn Hong Kong into a platform for tech development."

Science Park CEO Allen Ma acknowledges that while retail is important, "it is the technology, the intellectual property and the brain power that can give a city its competitive edge".

Gene Soo, a serial tech entrepreneur and co-founder of the StartUpsHK community, agrees. Ten years on, Soo views projects like Science Park and Cyberport as experiments we can learn from, but the time has come to try new models, he says.

"Could we better channel the money elsewhere? Figuring out how to kick-start Hong Kong's innovation is similar to launching a start-up. You can spend millions doing it and get nowhere or you can be nimble and do quick experiments to figure out an effective strategy to push innovation," says Soo, a computer engineer by training.

"We've been using the old models for too long. We need to move quickly and now."