How to improve OpenOffice’s blood supply

Writing in The Guardian, Andrew Brown wonders why OpenOffice, one of the flagships of the “replace Microsoft with open source” movement, is so bad. It is a question well worth asking, since in most discussions of open source software it is tacitly assumed that we are talking about nothing but a clash of business models and that the software that is on offer from both sides is essentially of equal quality – since if you have enough programmers then they’re bound to produce a product that works, aren’t they?

There is little dispute over the value of open source in tightly circumscribed projects with a well-defined function. If you want to create or display images in JPEG format, you would be mad not to use the Independent JPEG Group‘s software. If you want to encrypt your data link using the Secure Sockets Layer standard, you would have to have a very good reason not to use the OpenSSL implementation. In both cases the projects are run by engineers for engineers and the economics make sense: in exchange for giving away the work I put into identifying and correcting a bug in IJG or OpenSSL, I get everyone else’s bug corrections free of charge. If I need to move my software project to another kind of computer system altogether then I will find that someone has already done the work for me as far as JPEG or SSL is concerned, or at least that someone is in the middle of the same move and we can share the work between us.

As software manufacturers we would like to sell software with no bugs; or failing that, software where we can find any bug that is reported, understand it and correct it. Using open source components lets us do this without having to worry about the quality control of our suppliers – or even their continued existence (Cardbox seems to outlive everything).

But as users? As a user of Mozilla Firefox I have no interest in the fact that the source code is available to me and even if I found out how to get it I wouldn’t help me much. For the inner core who work on the project, open source is a way of transcending corporate boundaries and getting a critical mass of people to work together, but for the people on the periphery the only benefits of open source are the indirect ones: good software, or cheap software, or software that is immune to the whims and fashions of the marketing people.

A living software project, like a living body, has to have a circulation. From the centre outwards, there flow improvements, extensions, corrections of bugs. From the periphery inwards, there flow discontents, complaints, anomalies and reports of bugs. Without both halves of the circulation the organism dies. A small project, like a small organism, can orchestrate this flow by diffusion: insects don’t need hearts. But when a project gets big then something needs to pump that stuff around, and the best pump is money. If a user whinges, you sympathize; if a paying customer whinges, you jump.

This leads to an apparent paradox: free software stagnates when not enough people pay for it.

Paid-for free software sounds like a contradiction in terms, especially if you have listened to too many Stallmanite zealots; but it is the reality of what is going on. Relatively few people use Linux; many more use Red Hat Linux, Debian Linux, Mandrake Linux, paying (one way or another) for someone else to sift through the various open-source tools and projects, find those that work, put them together, and make them easy to install.

Brown says “[If OpenOffice 3.1 is a world-beater,] it will be because large companies such as Sun, Google, and IBM have decided that open source is the cheapest way to gang up on Microsoft, because it means they need spend nothing on support.” Brown is right that big companies are driving the big open-source projects, but the rest is fantasy. What these companies want is what we as software manufacturers also want: a product that is under our control and where we know that we can correct any bug that arises or add any new feature that is needed. Far from “spending nothing on support“, the whole point is that open source frees us from the costs of proprietary software development and lets us spend everything on support: it is what differentiates us from our competitors who are selling identical software to us. And our customers don’t have to worry about us taking a weird strategic direction or just going bust, because even if we do then there will be plenty of other companies around who can support “our” software.

So what of OpenOffice? OpenOffice has a difficult task because it is “big” in a sense in which even Firefox and Linux aren’t big: the job of a browser is simple and the job of an operating system is simpler still, but modern office software is exceedingly complex and has a rich set of interactions with its users. But the real reason that OpenOffice has trouble fulfilling this task is that not enough people are buying it, so there is not enough money to drive the all-important “venous circulation” that carries complaints and bug reports back to the developers with enough force to get them corrected.

On a corporate level this will change. The more open-source software that Sun and IBM sell, the more incentive they will have for improving its quality and the better OpenOffice will become.

But what’s more interesting is how OpenOffice could move forward even without the involvement of the big companies. The “venous circulation” problem presents an opportunity to design a robust circulatory system that will let even the largest open source projects grow and thrive. Suppose that Andrew Brown identifies his five most irritating OpenOffice bugs and offers $1 for each of them to be solved. Suppose that thousands – eventually millions – of OpenOffice users sign up to the “$1-per-bug” project. Who can doubt that the oldest and most irritating bugs will get solved very soon indeed?

Microsoft is facing the same problem and it has deliberately strangled its own circulation: short-term survival but long-term suicide. Microsoft has not solved the problem of distinguishing bug reports from support requests – it is probably insoluble – and it can’t afford to give support free so it has to charge for it. Charging for support suppresses the vital flow of information, complaints and bugs without which any software project must eventually stagnate and die. If you don’t have a support contract then it can cost you as an individual $200 or more just to talk to Microsoft. They say they’ll refund that if it turns out to be a bug but that still isn’t much of an incentive for a user of Microsoft software to report anything. You could imagine Microsoft tweaking the figures to increase the incentive (“if it’s a new bug then we’ll pay you $2,000”) but the numbers are so big and the multiplier effect of those millions of users so enormous that they simply daren’t do it in case it goes wrong.

The open source world doesn’t have this kind of money worry. It can experiment with all sorts of ways of managing information flows, of sifting out the real bugs from a mass of misunderstandings or simple user errors (but even a common user error is valuable information for a developer). Above all it can try different ways of injecting just enough money into the system to allow for a strong circulation and healthy growth. The actual money mechanisms are all there already (look at PayPal) – all that’s needed is imagination and inventiveness and a willingness to experiment.

Advertisements

4 Responses to “How to improve OpenOffice’s blood supply”

  1. Andrew Brown Says:

    ” Suppose that Andrew Brown identifies his five most irritating OpenOffice bugs and offers $1 for each of them to be solved. Suppose that thousands – eventually millions – of OpenOffice users sign up to the “$1-per-bug? project. Who can doubt that the oldest and most irritating bugs will get solved very soon indeed?”

    It’s a nice idea. But it has been tried several times, and never worked. I think Mark Shuttleworth had a bug bounty on OOo for a while. There seem to me to be two problems. The first is that most of the bugs are in the bits of the program that 90% of the users don’t touch. So the pool of potential buys is much smaller than it looks. The second is that there just aren’t enough developers who know their way around this huge and forbidding codebase.

  2. Leslie Satenstein Says:

    Andrew,
    In some aspects I agree with your suppositions concerning Open Office.
    However, having lived through the introduction of the PC, of the transition from DOS to Windows 1.0 and upgrade by upgrade to Windows XP, I would have to say that OpenOffice is at the Windows 2.0 level. Unlike your pessimistic belief that it will hardly improve, I see a great future for Open Office, and similar open products.

    My late father can recall when the only book one could buy was hard covered. Then we progressed to the softcover pocket book, that sold at 1/100th the price. Today, we read more (for free) from the net, then elsewhere.

    Due to the internet, which Microsoft preceded, that company, at it’s start, had to go it alone to manage and repair bugs. Today, with the net, the world is full of bug-killers.

    In short, the internet, and currency differences are driving the Open Software market, not to mention what satisfaction to one’s ego is obtained by working on the improvement of a world class product.

    A closing question. “Why are there artists in the world who produce for joy and hobby? Do we only want those who produce for profit? Why then do we do art for “joy”? ” In my mind, working on a major software product is “soul food”.

    My bottom line is, there is no stopping Open Free software.

    Leslie in Montreal.

  3. cardbox Says:

    Andrew –

    Thank you for your comment. I don’t know if you noticed the Slashdot discussion. To an outsider like me, the main point that stood out was that the “huge and forbidding codebase” was indeed the biggest problem. Like the other huge programming projects around (not least Microsoft Office) OOo seems to be a massive lump of interconnected code, and correcting anything in it is likely to break something else. Leaving the zealots to one side, the other overall impression was that an open-source project has to be born open source, both to ensure that it has the right structure and to ensure that it has the right community backing it.

  4. tkTim Says:

    Bugs vs. Features

    Brown made some valid points. I’m not sure buggy is the correct term to be using in all cases. The word wrap problem may or may not be a bug. It could be just a feature that OOO doesn’t have. I seem to remember when MS Excel didn’t do word wrapping for cell notes either. When the cell notes (comments) feature was first added the programmers didn’t think anyone would write notes long enough to warrant word wrapping in a cell. They found out they were wrong. For a long time Quattro Pro had a feature where you could color the tabs in a spreadsheet. We used MS Excel at work which didn’t have this feature. With every updated we would wish for this feature. It took them for ever to offer this feature, seemed like years. In fact there was a lot of features that Quattro Pro had that MS Excel didn’t have for the longest time. Corel took over Quattro Pro and removed some of the features to make it more like MS Excel. After complaints they added back some of the removed features. None of these things are really bugs, it just there way of setting up the program.

    One thing that could speed up new features for OOO would be for someone to publish a side by side comparison of features compared to MSO. One could even rate which features should be added or updated first.

Comments are closed.


%d bloggers like this: