I retired from personal blogging in July 2008.
But you can find me over at http://blog.xero.com.
I enjoyed speaking at the launch of the Auckland Centre for Software Innovation (CSI) Extenda program tonight. Extenda provides ICT companies with the understanding and tools needed to exploit Research and Development (R&D) activities in their business.
There was a good crowd in attendance and nice to catch up with friends and colleagues across the industry.
I started off comparing R&D for software companies against the inputs that traditional manufacturing companies require to create goods and services. Traditional companies require plant, materials as well as R&D. In software we don’t really require significant investment in plant or materials. The raw materials for software is R&D. In software:
R&D = Talented People + Time
Another contrast is the relationship between costs and revenue. In traditional businesses:
costs + a reasonable margin = revenue
In software there is the potential for almost no relationship between cost and revenue. For example, is there is no relationship between what the Microsoft Office team costs compared to the worldwide revenues they create.
Investing in R&D provides the opportunity to create this desirable dichotomy.
In New Zealand we seldom give ourselves time to do R&D. You cannot separate R&D from investment.
This lack of investment means that we do not invest in R&D to create digital assets, rather we fall back to capital-light service type businesses where our valuable talent is deployed to simply charge time. This is an inefficient use of our most precious natural resource.
The bottom line is that as an industry we are inexperienced in R&D. The Extenda program provides an academic framework and peer discussion opportunity to accelerate your R&D program. There are only 10 slots and at least 4 have gone so register quickly if you are interested.
http://www.csi.ac.nz/services/extenda/extenda-registration-of-interest
A bonus of my evening was meeting Falafulu Fisi - a frequent commenter on my blog who just astounds me and the Xero team with the depth of his knowledge. Great to meet you Falafulu and I hope we see you in Wellington soon. I hope you start your own blog because we all enjoy hearing what you have to say.
Good to catch up with Mark and Owen on the way home doing the late night Auckland to Wellington shuffle. (Fortunately I’d already learned not to book that last Qantas flight.)
blog
Thanks Rod, we enjoyed hearing your comments on investment in ICT R&D http://www.csi.ac.nz/news-archive/Extenda_launch_Rod_Drury_comment It’ll be interesting to track how your proactive investment approach to Xero impacts your business growth long term. We’re working with several companies right now supporting them making the pitch to the board on taking a similar approach, so having your example as an exemplar is very useful!
“In software there is the potential for almost no relationship between cost and revenue. For example, is there is no relationship between what the Microsoft Office team costs compared to the worldwide revenues they create.”
Well, that is helped by being in an monopolistic position and abusing that position. In a market economy one would expect the costs maintaining your software to keep ahead of your competitors to reduce your profits to something more economically reasonable (if not personally so!).
But I do agree the R&D investment in software is relatively cheap. Hence the explosion of ideas and sites see, the seemingly huge “overnight” successes and the string of dot com bombs. It is exiting and it is worthwhile participating.
A very good related article from Simon Hendery articulating the image problem in ICT, which Extenda takes one approach to addressing.
http://www.csi.ac.nz/news-archive/Extenda-lifts-ICT-image/
Don’s comment “Well, that is helped by being in an monopolistic position and abusing that position.”
Is missing the point. Back in the early 80’s MS wasn’t a monopoly, thats something that they’ve built. They initially did this thru innovation - taking a crappy system called dos and making the GUI real - thus having world wide appeal as you didn’t have to be a geek to use a computer anymore. (i’m not old enough to know the details and have heard they copied MAC)
Since that point, they have have scale, and that scale means their RnD is leveragble accross the many hundreds of millions of existing and potential customers… ie the digital asset
Hi Paul,
A good way to watch Microsofts rise is the for TV movie “Pirates of Silicon Valley”
http://www.imdb.com/title/tt0168122/
Not sure where you can get hold of copy.
My $0.02 is that R&D is desperately needed in NZ and I believe it can lead to the next, google, youtube, ebay … etc.
The risk to rewards of these type of businesses is fantastic!
Rod said…
A bonus of my evening was meeting Falafulu Fisi.
Rod I think that it was more of a bonus for me to finally meet with you. I enjoyed the whole event especially your talk and well done to the CSI organizers. It was good for us who are perhaps wannabe (or aspiring to be an) entrepreneur like yourself to attend such an event. Also, it was a good event to network with other people in the software industry.
Thanks for introducing me to Mark Donaldson of Phitek, and I had a chance to chat with him about his noise-cancellation-based technology, where he was happy to give me some general description of the product. We briefly chatted about signal processing algorithms, and I said to Mark that I will send him information regarding some recent advances in noise-cancellation algorithm researches (considering I am not an expert in noise-cancellation design, although the mathematics is the same one that I am currently using, but different domain of application though) that I have come across in the literatures that might be of interest to his R&D.
So, this late afternoon, I sent Mark 3 papers relating to 3 different algorithms on noise cancellation for his R&D team to take a look at. Those papers are already available in my huge (local PC) archives of publications in different computing topics (mainly in numerical computing), were sent directly to me from their respective original authors, in which is something I usually do, ie, request copies of their papers, if they’re of interest to me and also request their codes they developed to verify the claim in their published algorithms. If such codes have not yet been developed by the author(s), I usually developed one based on the published algorithm, and then send it to the original author(s) to verify its correctness. One can’t always read the mind of the author of an algorithm regardless if the procedures & steps are very clear & more detail in their descriptions from their paper of how to implement the algorithm in software codes. So, it is vital that one always has to double check with the original author about an implementation to verify the correctness (expected output from a certain input). Sometimes, they even correct it for me, and then send the modified version back to me. Authors are most of the times, to collaborate with anyone who wants to implement their algorithm.
I have also mentioned to Mark, that perhaps he could try and make contact with Prof. Nicola Kasabov, just to explore a possibility of collaborating with Nic’s R&D team on noise-cancellation algorithm development. Nic and his team at AUT KEDRI have published a number of papers in noise-cancellation algorithms. The KEDRI team’s papers, some are freely available online to download. In fact one of Nic’s team member, Dr. Richard Kilgour ended up working for Navman because the KEDRI team did some researches in Analysis of Noise Conditions for In-Car Speech Enhancement, which involves noise-cancellation to be able to enhance the speech. I am not sure whether Navman proceeded from proto-typing such system to fully developing one for commercial use. Perhaps some one who is a reader of this blog knows (Navman user) if Navman has such a product or not.
I have known Nic for about 6 years in which he has persuaded me in the past to move into his incubator at KEDRI, where he would provide me with a space. He knows that one of my computing interest overlapped with their (KEDRI team) computing interests in the area of Soft-Computing, where we all subscribed to the BISC newsletter, but Nic holds a special chair for the BISC Bioinformatics Special Interest Group. Also Nic is well-known worldwide on who is who, in the area of Neuro-Fuzzy computing. There was an international conference for BISC researchers held at KEDRI hosted by Nic in December 2006, which I booked to attend, but missed it due to family reasons. I told Nic then that Falafulu is not an entrepreneur, but time has changed, and I am attracted to the idea of being an entrepreneur so that was the main reason I attended the Extenda program last night to at least explore or find out the how to become one question. I am just mentioning Prof. Nic Kasabov’s name. The reason, I am mentioning Prof. Nic Kasabov’s name here, is that perhaps there are software companies out there including those that attended Extenda program last , who would want to explore any potential R&D partnership with such people as Nic and his team at KEDRI. The same thing applies to the Auckland University CSI. I encourage our software industry to find out about, how Universities R&D can help. You might be surprised at the results of a project scoping done by such R&D driven centre such as UniServices or KEDRI, perhaps something you never thought of before or even contemplate on.
I would like to say that I am interested in reading and commenting here at Rod’s blog. I am not self promoting myself when I make comment here, but I love to share what I know (to be honest I don’t claim to know everything) with Rod’s readers, because I know that a lot of these readers are software developers like me. Rod, I will leave blogging to you and others. I will just be a reader of other people’s blog.
Rod said…
I hope we see you in Wellington soon.
Yep, definitely. I will let you know (thru Andrew), when I am coming off from my current contract so that I can check out my calendar for a suitable time to meet with the Xero team.
Interesting stuff Rod, I think at the heart of it is that for a long time there has been no tangible R+D incentive. The tax credits the goverment has proposed are a good start but there’s still the problem of overcoming the short term fear that most businesses have of not backing the right R+D project. This fear means ultimately means nothing will get done and the business will chug along only focussing on the next bottom line. Once software companies get over this fear and start taking advantage of the opportunities given by this intiative with academioa and the new tax credit regime only good things can happen.
Well done to you, Nick Jones and your CSI team in organizing the Extenda program. I have a question because I am curious? Does CSI involve only the Computer Science and Electrical & Engineering Departments? I assume that it is the case here, as it is stated on your (CSI) website.
Just a thought, whether CSI wishes to or not, but I think it would be a good idea that CSI extends its collaboration program to invite the followings to participate in your program such as shown in the list below. I will also explain my reasons.
#1) Dr. Vojislav Kecman and his software team at the Mechanical Engineering Department. Kecman is the Head of Dynamics & Control Systems Group at the Mechanical Engineering Department.
#2) Prof. John Butcher and his scientific computing group (numerical computing) from the Math Department.
#3) Prof. Ross Ihaka and his statistical computation group. Ihaka and his team are very active in maintaining and supporting the R Project , which is a popular open source software in statistics.
Prof. Peter Hunter and his software modeling team from the Biomedical Engineering Department.
What I am suggesting above, is an opinion which I believe that the participation of such groups from within University makes you truly a Centre for Software Innovation, because you’re not only cater for business software but far reaching out to the engineering & scientific domain.
Vojislav Kecman told me the last time I talked to him at the Engineering School that his group is independent from the Electrical & Computer Engineering Department and also they don’t collaborate with the Computer Science Department. However he does collaborate with Prof. Allen Rodrigo of the Bioinfomatics Institutes, ie, Kecman and his team develop algorithms to be use in DNA sequencing & gene discovery by Prof. Allen Rodrigo’s researchers at his Institute. I am aware that there are some NZ companies which specialize in automation & control systems software and this is really an area where Dr. Vojislav Kecman and his team can help out CSI in.
I used to participate in Prof. John Butcher’s numerical computing group’s regular workshop but I haven’t attended in the last year or so, due to my busy working schedule where I have had no time to participate recently. This group can model & develop algorithm for anything really, from fluid dynamics, chemical kinetics, non-linear dynamical feed-back control systems, and many more. The kind of algorithm development that Butcher & his team are doing are complex stuff similar to what Prof. Peter Hunter from the Biomedical Engineering are doing with his team in developing cutting-edge simulation software in their Physiome Project. This software brings advances in diagnostic medicine never seen before. The overview of the Physiome Project is found here.
The type of algorithms that Butcher and his team are developing all the time is quite similar to the ones that Hunter and his team are using or developing and these are stuff in differential calculus, stochastic differential calculus, matrix algebra and so forth. I am aware that there are a few NZ companies who specialize in medical informatics software. In case these companies want to move into the development of medical diagnostic or medical decision support systems with special integrated hardware, then Butcher’s group and Hunter’s group could offer their expertise in that area to CSI.
Ross Ihaka and his team at the Statistics Department can offer their expertise in the development of statistical computation algorithm for automated predictive modeling. These guys are quite good at what they are doing. In fact the R Project originated with Ross Ihaka and the team at the Statistics Department in the 1990s. Now this software is widely adopted by researchers as well as practitioners in the industry. Dr. Robert Davis from the Business School (Tamaki based) specializes in M-Commerce for Measuring Interactivity, Advertising Effectiveness , Impulse Purchase Behavior and other areas. I met up with him during one of his project in trying to quantify Advertising Effectiveness via software platform for TV text-messages as in NZ Idol , Sporting Events, etc, by using predictive modeling. He wanted to see if I could help him out by applying data-mining and machine learning to his predictive modeling problem. I looked at his data and I told him that the problem is not to be solved using machine learning and data-mining (not enough variables or data) but statistical predictive modeling is the right way. So, in the end, he hooked up with his contact at the Statistics Department where they developed the core predictive algorithm for him and coded .NET by Robert Davis’s students. The software company they worked with is a local company and Robert Davis did this via UniServices. This company markets their software to the US and they want to expand to Europe. If such a similar company wants to participate at the CSI program in the future, then definitely Ross Ihaka and his team can lend their expertise is statistical modeling if they are part of CSI.
It is just a thought, Nick Jones and I believe that wider the spectrum the area of expertise at CSI, the better for our software industry of all sorts (business, engineering, scientific) to become engage with CSI in terms of R&D.
Finally, I don’t believe that I had a chance to talk with you at the Extenda program the other night, but I did briefly talk to both Santokh Singh & John Corey. Perhaps, next time then.
Cheers.
It seems like such an obvious connection to make. Universities need funding for research projects plus industry practicum for their students - and technology sector firms need a pipeline of good ideas to commercialise research that might otherwise lie fallow. It’s win, win.
Auckland University and Uniservices are to be congratulated for their successes in matchmaking “town and gown” over the last few years. No doubt CSI will extend this success further. And I hope their colleagues down this end of the island will also take note. With the “centres of excellence” proposed in the Wellington regional economic strategy there is now an opportunity to build some meaningful collaborative capability that is currently lacking.
“This lack of investment means that we do not invest in R&D to create digital assets, rather we fall back to capital-light service type businesses where our valuable talent is deployed to simply charge time. This is an inefficient use of our most precious natural resource.”
Very apt! The one tweak I’d make is there’s more problems than just that. Its not just investment; I think its also hard to launch products here because there a few NZers to buy your products (small, remote population), which means you have less early adopters, and also less wealthy people. This means its hard to get to a tipping point, and get a mainstream customer base and make profit. I wonder how all fares in Australia, given its a halfway point between us and the US…
I’m very thankful for the Google Summer of Code programme that’s given us ten smart programmers to build up our SilverStripe product at little cost (i.e. time only); its been marvelous to see in the last week all of the work they’ve done (except for the fact I’ve been up past midnight for several nights updating our website to give updates on them!)
Its going to be interesting in a week when over in the states to constrast the NZ way with the US way… will certainly be bringing back some knowledge…
I’ve spent most of my working life working for startups in the US (before I moved back to NZ).
My experience has been that medium to long term success at that end of of the biz spectrum has a lot more to do with luck. You have to have a bright idea, the willingness to work your ass off and people who will plonk down some VC money and experience to get you to there from here (my one attempt at self financing was really tough). But in the end it takes long enough to get to an real end product that the world has changed around you..
In the end maybe you’re flavour of the month when you roll it out (even that doesn’t last, trust me), maybe you’re not - it’s a crap shoot, anything that takes a year or more in such a fast moving world are a risk and most that don’t probably aren’t worth doing.
There’s a maxim that out of 10 startups 1 flies high, 2-3 muddle on to some success and the rest die those are the odds what the VCs plan on getting their money back on - I think there’s an expection for the guy with the totally brilliant no one else has ever thought of - but those are probably 1 in 100.
Now I’ve only been back in NZ for a couple of years and don’t live in Auckland or Wellington so I’m not really plugged in - but I do kind of get the impression that people are kind of hoping for the 1 in 100 sort of startup (instant wealth) rather than the 1 in 10 sort (hard work), probably that’s just human nature - and I know there probably isn’t a lot of local VC money floating around, maybe KiwiSaver will loosen some more up in a few years.
Finally while there have been some great startups come out of academia in the past (Sun, Google, etc) my experience is that most involve people with more real-world experience (and I’ve hired my share of fresh Stanford PhDs in the past into them, some of who worked out and some of who didn’t really function in the startup world).
Very interesting conversation. But is there support for productising a good R&D idea? I know a lot of smart people with good ideas in IT, but the point of struggle is how to bring a product to the market and let the market know. Smart IT people are not set up (IMO) to perform these functions. It seems the big problem is to get smart IT people in touch with smart marketing/sales/legal type people, without big chunks of money being involved.
Im struggling with this issue myself, being a smart(ish!) IT person with no marketing background. What support is available for this, essentially the opposite situation?
regards
Greg
I guess it depends on your product, if your target customer is yourself (other IT people) you probably know where your market can be found and are halfway there - if your market is the rest of the world (I currently build TV stuff) you need someone who knows how to get there - maybe contacts in the retail chain, marketting to get your big idea in front of the public and to make them understand why they want it.
In my experience there’s good and bad VCs - good VCs want to make money, and want you to too, they know where to find the CEO you need, the marketting people you need etc etc really smart VCs seem to keep these people on tap (good engineers too) maybe bring them in house between gigs to help to look at upcoming projects - bad VCs of course just want your money ….
It’s all about dialogue. Smart ideas thought up in academic institutions are nothing without the wherewithal to take them to market. A poor product with a good team of marketers and some strategic visionaries will always outperform an exemplary product without there key features.
The incubators set up near the universities are a good start - especially when some of the economic development agencies, NZTE, ConnectNZ are involved.
It’s a huge area of concern that good ideas are being born in NZ but we don’t have the business accumen to execute them
Paul I don’t agree with those 1 in 10, 1 in 100 crap shoot type stats for start ups.
Looking in the rear view mirror you can say 1 in 100 makes it. But it is not a crap shoot.
Start ups fail for many good reasons. Lack of cash, lack of talent, dumb idea, too early, too late, lack of market etc. etc.
Any business that I do I’m going make damn sure it works or if it is not making it - kill it quickly.
The businesses that fail were always going to fail. The ones that make it have enough right and luck. Most luck you make yourself.
Irony loves me, the captcha word is “relax”.
There is very little support here for software research. I self funded (off the sale of other intellectual property) the development of an h.264 implementation that first went end-to-end live in January 2003. At the time it was the only implementation able to encode live on generic hardware. I realised my strengths lay in software, not business and hence started working with [deleted because the Internet is a small place] who made a lot of promises and delivered no results.
Of course, being an entrepreneur, mere failure was not an option and as a capital raising exercise I took on some other projects, took on some other people and finally had a badly underfunded (and by now late) push for the finish line. It did not, of course, work. It was the wrong finish line too, I should have taken the nearer one :)
I am still bitter about it, truth be told, and while anger at some of the players has subsided into a better understanding of what poor quality people look and sound like, my belief in the NZ business community’s ability to commercialise innovation has only gone down. That being said I am stoked about both Phitek and Syft.
I now believe business should start with the dollars. Sorry, idealists, you lost me.
But how to get to those contacts? If I am a smart software developer who develops a useful product for a given market, how do I get contacts in that market?
whats the link from “I’ve got a cool prototype” to “help me get this to release 1.0″? I can’t see where this link turns up in the scheme of things, and I know there are a lot of motivated IT (development) people out there with access to a compiler and a lot of passion.
seems a missing piece of the puzzle, unless Im missing pieces of my puzzle! And I don’t believe its a crapshoot either, although the “bringing product to market” phase seems a lot like that to a non-marketing, non-sales type person. D
Greg you need to network and find complimentary people who fill in the bits of the puzzle you don’t have. And that means networking outside of your own gene pool. NZSA events in Auckland are an easy place to start but look for conferences in your area of specialty.
There is a myriad of organizations that help. From incubators to NZTE. While they won’t do it for you they should be able to point you in the right direction.
Rod is right - there are a plethora of organisations that can help find the people to fill the missing parts of the jigsaw.
We all just need to be more willing to seek, and accept, help and guidance
Thanks for that guys. My point really is that there is a lot of expertise and passion floating around outside of universities, but not necessarily an obvious route (at least to the uninitiated) to productising this energy. Not all software developments have to be taking on the world, even niche markets can be lucrative and can go on to fund other things.
Who would do anything else? Just bring passion and a compiler.
:-)
Rod - I guess we have to disagree - my experience has been that the world changes around you while you work on your product - a year is a lifetime, you honestly don’t know if there’s a guy in the next building, or across the world working on the same thing, or something better.
I used to design graphics cards - we had 2-3 strong competitors, and could never tell if a new product would be a boom or a bust ’till it hit the market - we bet the entire company on a new chip packaging technology once (and won big time! - the downpayment for my house) - I see ATI/nvidia in the same race today a decade later
People’s ideas of what’s hot change just as fast - web 2.0 may be flavour of the month today and get all the press, tomorrow it’s ho-hum another online spreadsheet ….
My point is that you honestly can’t judge when you start what things will be like at the end, you put your money down, work hard and take your chances - that’s the 1 in 10 startups make it big thing (and remember that’s VC funded startups - by people who ought to vaguely be able to pick winners).
The 1 in 10 concept for startups is an old idea, one I’ve heard expounded a number of times, in print and directly from VCs.
The 1 in 100 thing is my own take (and just a number pulled out of my arse) on the idea that there are genuinely new revolutionary ideas out there, ones that will make new markets - ideas that if you have them and get to own and develop them you really will make it big - they are few and far between and not always obvious - there’s probably only 1 or 2 of those show up world wide every year, and often it’s not obvious for a while (google’s a great example).
Human nature says that you know that your idea, your startup is the next best thing - looking past that is hard, I’ve seen VCs get sucked in too - the .com boom/bust was an extreme example of this - in my family we now have a rule about not counting chickens - the first 2 times we became paper millionaires it didn’t stick …. and we didn’t even work for .com companies.
This is not to say that people shouldn’t do startups - you should - they’re fun, cutting edge, with amazing energy - I wouldn’t work anywhere else - I do want more of them in NZ - but failure is normal, especially if you risk big for big rewards
Interesting conversation, everyone.
I have to agree with Paul Campbell’s comment that “while there have been some great startups come out of academia in the past (Sun, Google, etc) my experience is that most involve people with more real-world experience”.
I couldn’t agree more, Paul - the way I see it, you cannot learn to drive a car in a classroom, and commercial software development is no different. Efficient design, good error handling, true adherence to object-based principles, robustness and so on are all born of experience and 100 hours weeks, and there is just no substitute for experience.
I am a developer with 20 years of experience.
I went through the whole experience of developing my own application, with minimal or no capital.I went through a succession of business “partners”, all of whom invested little or no money, and expected a return in the short term. Basically, the whole process was bruising for me….and in the end, I upped and left because I had just had enough. I bought my product over to the UK and have had considerably more success with it that I ever had in New Zealand.
I now work in London, and I am extremely well paid as a contract developer.
I have always been told that Kiwis were exceptionally good software developers ( I am old enough to remember the Porter project ) but I have always taken that with a pinch of salt until I came over here. Having worked for 2 very large worldwide businesses, both with large dev teams, I have to say that I now consider that it’s true, I find that the development standards and general level of capability over here are very, very weak, and it frustrates me.
And finally, a question for you all, particularly Rod.
I have read several comments made by Rod about the fact that developers should work to create intellectual property and add capital value to their businesses, rather than just billing hours, or words to that effect.
I agree with him, but if you look at my own experience, exactly how do people like myself do that in New Zealand ?
I spent five years developing my own app, with extremely long hours, and I earned a fraction of what I am now earning over here. I have a family to feed and in the end I couldn’t ignore the need to earn a decent income any longer.
Unless developers are now paid considerably more than they were when I left, returning to New Zealand to work as a developer just doesn’t make any commercial or economic sense. I would absolutely love to return to New Zealand but development jobs like I have over here just don’t seem to me to exist in New Zealand ( other than perhaps in the Xero project, which I am watching with interest, but they seem to have some pretty bright developers working there already ).
So what to do ?
Cheers everyone
Paul, I think your comment:
“you honestly don’t know if there’s a guy in the next building, or
across the world working on the same thing, or something better”
is a really interesting one and hints at something important in the
focus of many startups.
One of the thing I observe about Rod’s blog and its associated
followups is that the focus is on software products which “anyone
could build”. This is not meant to sound dismissive - what I’m
getting at is that given the key idea for the product, most good
software developers could probably arrive at a reasonable
implementation.
This makes it incredibly difficult to compete. You’re worrying about
someone else coming up with the same good idea, or when you launch,
seeing your idea and racing through a mark-2 with a better
implementation - there are a lot of talented software developers out
there trying to find a niche. The development tools market is an
extreme example of this, as pretty much every software developer
understands the requirements right off the bat.
I’m really interested to hear of NZ IT startups building products
for complex niche markets - those which take years of industry
experience to really understand. It’s my belief there are a huge
range of opportunities out there as there are so many very poor
systems around. For products like these I believe you’re far less
likely to be beaten to the punch, as there are so few teams capable
of doing a reasonable job.
From FF …
Nick Jones, I have another suggestion. If you’re already planning on
doing something similar, then please ignore this post, but if CSI hasn’t
thought of it, then I would encourage CSI to explore it further.
I believe that if CSI produce some kind of an online newsletter and
distribute it to all local software companies on a regular basis, so
they can digest the summaries and see if there is any concept covered in
that newsletter which might be of interest to their product development.
If a company finds a concept that is directly applicable to their
software but perhaps they have never heard of such concept or being
aware that such thing even exist, then they would definitely want to
seek help from CSI if such capability is beyond their domain of expertise.
I believe that software companies aren’t going to come knocking on CSI
door, seeking help unless:
- they’re aware of something new in which they lack expertise on
- perhaps they try to implement something and on doing so, they run
into a brick wall and that might prompt them to seek help.
- and may be more…
As a result many companies wouldn’t want to approach CSI because they
think they don’t need any external help. In most cases, it is true,
however in other cases, some companies just completely missed out on the
latest advances in a specific software technique, in which they’re
unaware of , most likely because they don’t read computing literatures
that are relevant to their domain. This is the advantage of a newsletter
is to inform the industry of what is the latest (which if implemented
via collaboration with CSI), has the potential to improve the
functionality of their software product(s) and make them very
competitive in international market. The point that I am trying to make
here, is basically a follow on from Prof. John Hosking’s Herald comment
which I have cut & pasted below:
From the Herald Article by Simon Hendery that you linked to:
John Hosking, professor of applied computer science at the university
and a driving forces behind Extenda, says one problem is that senior
technology staff in New Zealand businesses tend not to have
post-graduate or often even degree qualifications. As a result, the
companies they work for, while often successful at selling a single
idea, can struggle when it comes to continuing to grow through innovation.
“Although they [the technology managers] are bright and innovative, they
haven’t been exposed to research methods as you would through doing a
masters or PhD,” he says.
“We’d like to open their eyes to what research can do to make companies
grow quicker. Obviously we’d be interested in helping them do that
research - we’re quite open about that - but the name of the game is to
raise a research culture within their companies, and allow them to grow
quicker, be more innovative and avoid this one-product wonder trap that
they get into.”
Here is what I think. Since, Auckland University do subscribe to most or
all of the computing & engineering related journals, a summary in a
monthly newsletter will have an issue dedicated to say topics that are
applicable to Knowledge Management System and Intrusion
Detection Systems and the next month will perhaps topics that are
applicable to CRM , Document Management System or
something else , etc:
My example is as follows. Suppose that the July newsletter covers topics
that are applicable to Knowledge Management System and
Intrusion Detection Systems. Yes, I know that we do have local
companies that specialize in those types of software products.
#1) Using
Case-Based Reasoning to Improve Information Retrieval in Knowledge
Management Systems
#2) ROUGH
SET THEORY’S APPLICATION ON INTRUSION DETECTION BASED ON SYSTEM CALLS
#3) Network
Intrusion Detection Using Wavelet Analysis
The newsletter could just reprint the abstract for that particular topic
itself or perhaps summarized the whole document as to be friendly,
readable & understandable by the technology managers of companies. That
is to avoid putting too many technicalities into the newsletter, but
give a readable overview of why this new technique is better than the
current existing ones. This would make technical managers want to knock
on CSI door as soon as they finished reading such newsletter. In the
newsletter, it should also describe the weaknesses of the new technique,
just to avoid technical managers being thinking that the new technique
if implemented will conquer them in the world markets against their
competitors.
A company that develops knowledge management software but doesn’t know
that there is a better way or techniques (algorithms) to be developed
into their product which has the potential of making it world class,
such as Cased-Based
Reasoning (CBR). I doubt if any of our local knowledge management
software company is adopting CBR. A newsletter would make them become
aware of such thing.
The advantage of a newsletter will inform such companies, that there is
a better way for knowledge management system. I believe that if there is
such an online newsletter from CSI, the number of companies come
knocking on their door would increase over time.
Ref #2 & Ref #3, were references I sent last year to a local software
company that develops network intrusion detection systems. They weren’t
aware of those algorithms, but basically I offered to develop those ones
for them. They currently use Artificial Neural
Network (ANN), and their response back to me, saying they were just
happy with ANN and did not need anything new to be added. The use of
wavelets & rough
sets had been shown in recent publications that they outperformed
ANN in regards to network intrusion detection in certain types of
intrusion (not all). This was the reason I made contact with this local
company to show them the publications, because it wouldn’t hurt their
product if they are (wavelet & rough sets) being developed into it as
extra capabilities on top of their ANN based algorithm. The reason I
brought this up, is because if reflects directly what Prof. John
Hosking’s comment on the Herald, which he stated, that while often
successful at selling a single idea, can struggle when it comes to
continuing to grow through innovation..
Paul Cambell said…
I guess we have to disagree - my experience has been that the world changes around you while you work on your product - a year is a lifetime, you honestly don’t know if there’s a guy in the next building, or across the world working on the same thing, or something better.
I agree with your statement here Paul, but I think that is exactly what CSI is there for in the first place. They would suggest to you (vendors or start-ups) of what are the latest techniques that might give a competitive advantage if implemented. Sure, some guy might be working on the same product across the other side of the world, perhaps in Russia or Finland, and this person is taking advise from a similar institute such as CSI in his own country, such person might be on the upper hand compared to a NZ guy who isn’t taking advise from our own CSI.
I think that the whole idea of CSI is to open up the technical managers (or developers) eyes. In my previous comment, I gave an example of using Case-Based Reasoning (CBR) for development of knowledge management system. CBR models how a human recall things from its memory, basically it could be generalize as memory-retrieval-system. I am not sure if any of our local knowledge management system vendors have implemented CBR (or similar techniques - yes there are lots of them) in their product or they are just using the conventional query-type SQL. CBR beats SQL query-type by a wide margin in terms of the Recall and Precision capability of the Information Retrieval System. A system with higher Recall and Precision is much better than one with a lower value. Risk for finance companies are rated as AAA, AA, A, BBB, BB, B, etc, and the higher the rating, the lower the risk. It is the similar to rating of Information Retrieval System, the higher the Recall and Precision percentages, the better the retrieval capability of the system to retrieve relevant information it is exactly why I compare CBR & SQL here. One outperforms the other.
The chance that a software developer would have come across CBR is small compared to someone who have read it somewhere else (such as UNI library, Amazon book) or perhaps read(heard) it in a software newsletter (seminar). This chance is definitely increased if companies explore what CSI can offer them. Here is a fact. People only Google for info on things that they have come across or things that they already know. So, things that they don’t know or have never come across or heard before, will remain unknown to them, until they read some article or a newsletter that mentioned those things (perhaps some algorithms or techniques).
John said…
Efficient design, good error handling, true adherence to object-based principles, robustness and so on are all born of experience and 100 hours weeks, and there is just no substitute for experience.
Yes, this is true. When one develops something, that person must adhered to those principles you have stated above even when you design a CBR system or Artificial Neural Network system and those principles still apply. What I am advocating here, is that institute such as CSI would be able to help you out with things that you don’t know, which perhaps lift the capability of your products from the conventional way you have been doing it in the past.
BTW, I am not an academic, a software vendor nor member of CSI, but I quite like the idea of having such institute as CSI to work with the software vendors in our industry. I use the university resources, such as browsing the regular issues of their computing & engineering journals just to basically look out for new algorithms or new way of doing something.
I do have an advanced degree background, and the amount of research out there is staggering. Sorting the chaff from the wheat is also pretty difficult unless you have a lot of time (or a significant R&D budget - which is what we are arguing for).
I think the direction is encouraging, I also think that it would be a good idea to have regular meetings organised by universities that were published more widely, and were open (and advertised) to the public (eg: via the newsletter that Rod is proposing).
Its very easy in a university atmosphere to become isolated from the wider world, particularly in IT. I think it would be a positive step to have graduate students, especially PhD level, present their work to an interested business audience, and be particularly encouraged to think about applications outside of their thesis area (which are generally mind-bogglingly narrow). The students get practice at presenting, and the community get informed about what is happening there, and possible ramifications which could then be followed up.
I certainly hope we could move towards this, along the lines of the unlimited-potential free pizza evenings.
I wonder whether one of the problems with R&D in New Zealand is that we place a great deal of emphasis on ‘D’ and very little on the ‘R’.
Ideas arrive via serendipity and are forced into the pipeline on the assumption that they are good i.e. worth pursuing.
How much preparatory investigation goes into most software reseach to determine the size of the market for a product before the prototype is developed?
I once worked for an IT company that was bought by a publicly listed company (cheaper to own the business and make a profit from their services than to pay for them on an hourly basis). The firm got it into its head that their future lay in software development, rather than internet services - licensing IP was the way ahead. Good in theory but flawed in pactice. They lost focus. The products the were so pround of were not unique and were worse than some which could be got for free in the web (email client and a CMS). What remained of the firm was sold off swiftly by their parent company who realised the mistake they had made before too long.
The story serves to illustrate my point though. Had the business spent time with real people in the market they may have realised that their development time was wasted and could never be recovered - busy work.
I often hear that ideas arent worth anything until they get done.
The truth is that even when done most ideas aren’t worth anything.
If you are going to make a profit from your ideas they must serve an unmet need (equal emphasis on both terms). And don’t get me wrong - the market is mostly inarticulate - going out and simply asking them what they want won’t provide an innovation or insight that is going to make you rich. Who could have articulated their latent desire for an ipod when proto MP3 players were functional but uuuuugggggleeee (I call that imponderables - we can’t know what we don’t know).
So the issue of research is paradoxical. But, then, if it was easy…everyone would do it. If you want to excite worldwide interest then you might also have to go out into the world and have a look around at how people behave. You might be surprised to learn they’re not sitting around waiting for you.
John: my take on your comment about comparing Kiwi software developers and others - I moved to the US 20+ years ago and I saw the same thing - to me it seemed like a product of the way that my late 70s University undergrad education worked - my classes didn’t leave me with a lot of choices, I had to do a CS course in breadth (even, gasp! cobol) while people I worked with in the US had been allowed to focus and new some stuff really well but had gaps.
Now this is one data point and probably a bit of an overgeneralisation - but that Kiwi Jack-of-all-trades works well with #8 wire world view has done me well over the years
For any vendors (or developers) who would be interested in looking at solutions (or techniques) for decision support systems (any types such as knowledge management system, engineering fault diagnosis system such as aircraft, medical diagnosis system, help desk system, e-commerce product recommendation, etc), such as the CBR, which I mentioned in a previous post in this thread, might perhaps want to make contact with Dr. Ian Watson of Auckland University, because he is one of the top researchers in the world in this field (CBR). CBR is defined as retrieval of the closest match (similarity search) from memory (database) rather than exact match (as in SQL query). He is also part of the Artificial Intelligence that works in partnership with CSI. So, either contact him (Ian) directly or the alternative is to contact Nick Jones and CSI about the possibility of doing consultancy or R&D in CBR based technology for your business.
Ian has written 2 books (both very popular) on CBR based technology:
- Applying Knowledge Management : Techniques for Building Corporate Memories
- Applying Case-Based Reasoning : Techniques for Enterprise Systems
There is also a series of conference proceeding papers in CBR research which had been published in Lecture Notes in Artificial Intelligence (LNAI), and these volumes are available at Auckland Uni library (Main). I think that all our NZ University libraries must have them (check your local), since it is impossible for University researchers to conduct R&D without getting access to published papers of their peers in their fields which are available from international literatures.
CBR is based on k-Nearest-Neighbor (kNN) algorithm. Developers who are interested in developing CBR based system must look at the popular WEKA machine learning Java project from Waikato Uni, because there are 2 sub-packages for building CBR available (this and this). Some WEKA users had already made WEKA workable in a .NET environment, so .NET developers must check out the WEKA non-core contributed packages of how to do that. The sub-packages in WEKA are not ready-made CBR inference engine yet, since WEKA was developed as a knowledge-discovery system (data-mining) rather than retrieval system, so WEKA packages could be tweaked a little to make it capable of retrieving via memory-based.
If you want to play with a ready made CBR Java API inference engine, then request a copy of the US Navy Conversational Decision Aids Environment (NaCoDAE) from Dr. David Aha. You must sign a contract form with the US Navy about the use of NaCoDAE and there is no hassle in doing this. It is only a one page form. Just let David know that you’re from New Zealand, because he loves NZ. He has been here twice in the past for international conferences on CBR held in Auckland, which he co-chaired with Dr. Ian Watson. I met David at the last CBR conference that was held in Waiheke Island, Auckland, 2003. NaCoDAE is not domain specific so it can be used for any decision support systems regardless of the domain (law, commerce, engineering, etc).
This series that I have come across may be of great interest to Rod & Xero (or any who’s interested in the subject really). There are 6 volumes on Artificial Intelligence in Accounting and Auditing available for sale online. Volumes II, III, IV, V, VI are listed on that page in the link highlighted above. I haven’t read any of those titles, but they look very interesting. Of course, all the techniques discussed in those books are domain expertise of the AI group that works in partnership with CSI, so there is an opportunity there for Xero to explore these technologies with CSI. The content of Volume VI is the most interesting one, where it covers the use of Artificial neural network, Software Agents, Fuzzy System , Expert System , Rough Sets , Case-based Reasoning etc, in accounting & finance.
Perhaps Xero could just nitpick certain topics or case studies from those books and present them to CSI as interesting area for R&D.
This is the correct link for the volumes on Artificial Intelligence in Accounting and Auditing.