Here at OSS Watch we often get asked things like “which is the best open source twitter alternative” or “which is the best open source financial managment tool?” We don’t answer these questions, and for good reason. Instead we focus on helping people understand how to find and evaluate open source software solutions. In this blog I’m going to try and explain why we take this approach.
With simple things like microblogging we can point to Laconica with a resonable amount of comfort, but should we really be ignoring Open Micro Blogger simply because we have not used it ourselves, after all, it does focus on a stronger and more scalable back-end. What about the solutions we’ve not yet heard of? A quick search for “open source microblogging” turns up a fair number of projects that are worth a look, not to mention a promising looking article with the title “Top 10 open source microblogging services. Perhaps it’s not as simple as we first thought.
If it is this hard to identify “best of breed” open source solutions for something as simple as a microblogging service how hard would it be for a “Financial Management and Reporting” system? These are complex systems that need to be customised (or at least configured) for each use case.
Sometimes the enquirer pushes a little harder, often indicating that it should be possible to identify “best of breed” and that OSS Watch should do so. However, our position remains firm, we can’t evaluate everything out there, but we can help people identify and evalate the options in the context of their needs.
Lets consider the financial software market. I’ve used Compiere for Financial Management in the past. I was very pleased with it, but it’s not really a financial system, it is a Enterprise Resource Planning and Customer Relationship Management system that has some financial functionality.Would the CRM and ERP functionality help or hinder our requestor?
As well as Compiere I’ve used a pen and paper based double entry ledger system. It worked, requires minimal technical knowledge and is easy to source. Unfortunately it is error prone, requires considerable training and usually needs a real accountant to do magic periodically, magic you don’t want to be paying accountant hourly rates for when software can do it, perhaps GnuCash or jGnash would be a better option. However, these are targetted at personal finance not business (or academic) finance.
Then there are OpenER, SQL Ledger, TurboCash and all the others that are out there. These are all “best of breed” within their own domain. They all have strengths and weaknesses. The “best” depends entirely on whether their strengths are important to the user whilst their weaknesses remain hidden.
How did I come up with the (incomplete) list of financial tools above? Simple, I spent 7 minutes (yes I timed myself) searching the web and making snap decisions on the viability of the open source community built around the products. I rejected another four or five as they looked like unsustainable or immature projects. It is this skill that I think we can offer to people looking for open source solutions. The customer must then narrow down the choice by looking at the feature set of the product base. Once a short list is drawn up, OSS Watch can help evaluate the support options available for the project, that is we can help the customer decide if the open source solution fits their culture, e.g. is there a suitable commercial support offering or is there an active and supportive developer community that internal IT staff can engage with?.
Even when we explain this in some detail we find that some people still insist that we should be able to tell them what to buy. So, I’ll say, here and now, as loud and clearly as I can:
Procuring something without actually considering the functionality it will need is insane. It is almost certainly doomed to fail.
A software procurement specialist once told me “there is never enough time to do it right, but there is always enough time to do it twice”. I wish I’d listened to him then (I’ll not go into why, but involves lots of wasted money). At least I’m listening now. If you are ready to listen, you can stop reading and contact OSS Watch when you want us to help you evaluate some open source solutions.
For those still unconvinced let me try and illustrate the problem in a domain most of us know a little about.
Imagine you are buying a new car. All cars have the the same basic functionality and features, they go forwards and backwards, they have a steering wheel and they stop before hitting things (assuming the driver is alert enough to hit the brakes).
Despite all cars being the same (when considering base functionality) we don’t walk into a the local car dealership and say “what’s the best car” and immmediately buy the first thing that comes into their mind. Instead we find out which are the most suitable cars for our kind of use (driving to the shops, commuting, long distance travel, fun at the weekend etc.). We probably also read literature and reviews. Most of us would test drive a couple of our favourites. Eventually we would buy one.
It’s true that we rely on reviews and salespeople to help us narrow down the choice, but we have other methods for making the final selection, and, importantly, we would usually ask salespeople from each of the major manufacturers.
Anyone who watches car “review” programmes like Top Gear knows that even in the simple world of cars there is never a clear winner between two cars built for the same purpose. One will be better cruising on the motorway, the other may be better on country lanes. One will be cheaper to run, but the other mighter be cooler to look at, but beauty is in the eye of the beholder. It is not usually possible to say “this car is better than that car, regardless ofwho you are and what you are looking for”.
Sometimes it is possible to say “this car is just rubbish, it broke down 5 times in our road tests”, but once you get past obvious flaws that affect everyone you have to understand the drivers needs before saying “this car is the perfect car for you”. I’m 6′ 4″tall, I simply don’t fit comfortably into some cars, it doesn’t matter if everyone else in the world thinks it is the perfect car, for me it is rubbish.
[This is getting to be a long post so I won’t even start on considering maintanence options for the car once it is purchased.]
Since most people wouldn’t ask a really general question such as “what is the best car?”, why do people insist on asking OSS Watch things like “which is the best open source financial management software solution?”
OSS Watch can help you procure open source solutions, but we can’t tell you what to buy - sorry, it just is not that simple.
It’s a fair argument but I do think you are missing one point. When we want to buy a car, one reason that we don’t ask someone to *tell* us what to buy is because we believe that everyone you mention (manufacturers, sales people etc) has a vested interest in recommending one car over another. When there is someone in a particular domain who is perceived as unbiased then we do quite often seek a specific recommendation. Gong with your car example I understand that the car chosen by “What car?” magazine as their “Car of the year” certainly does have an impact on sales. So I guess that if OSS Watch is perceived as unbiased then it is natural and understandable that people will want us to recommend specific solutions even when we know we are not in a position to do so.
You make a good point Elena. I think my car analogy was not the most efficient. Another is when we consult a doctor. We are worried, we just want them to tell us what is wrong. They are unbiased, they are not trying to sell us anything.
However, much as we want an answer quickly we would rely on a doctor to tell us what they are sure of, not to hypothesise without the full information. We don’t want them to tell us we have cancer and start a course of chemotherapy unnecessarily. We expect them to wait until the test results are in and they are certain of the diagnosis. Then, and only then, do we want to go through the treatment process.
Doctors are the true unbiased advisors here, not the car salesman, the mechanic, the friend in the pub or What Car? magazine. OSS watch are the “doctors” of open source and we do not advise on a course of action until we are certain it is the right course of action.
One of the biggest hindrances to giving pat recommendations is that different users have different requirements. Your example of Compiere is a great one. You found it useful for your needs, but that doesn’t mean that someone who requires a more specialized financial package will find it a good recommendation for them. Of course, you can always use the approach of “This is what I needed, what I used, and why it worked for me,” which is certainly helpful.
That said, I really appreciate your post because for me, the “Top Ten” types of posts are starting points, not decision points. Most software in the enterprise space, in particular, is too complex and generalized to allow yourself the luxury of picking a package from somebody else’s list, IMO. Recommendations can help, but there’s nothing like taking it out for a test drive and comparing it to others you’ve driven.