OS Support and Wallpaper Paste

25 Jan 2011

In a perfect world, all versions of Lasso would support all known operating systems, datasources and all known versions under the sun. I also hope in the future it will allow various syntax options (e.g. writing in php and lasso syntax at the same time on the same page, so we can steal crappy code from other languages).

Unfortunately, the current realities are much more economically and logistically driven.

As a new LassoSoft group, we sat down and listed all operating systems on all platforms in various bit flavours, as well as the requirements to create builds. Based on a "hopeful" figure of 4 days per OS, we realized that the process of supporting every common OS could take 6 months.

Not only that, but there are basic hardware and software requirements for development and testing;

- A VM, or in Mac's case, a unique machine
- A "clean" copy of the OS
Setup time = .5-2 days per bed, per person

This needs to be set up and perfectly duplicated for each team member testing and developing on this platform. Currently this is 4 people (2 developers, 2 testers). As one cannot develop Lasso 8.5 on the same machine as Lasso 9 (given the library issues), each piece of hardware/software must be duplicated for each major build we wish to do testing with. So for one developer, for one independent OS version install, we need a full development and full testing suite (at minimum).

Setup for 8.6 OS building so far (which had to be started on each OS from scratch) took a total of 4 weeks of developer time for 3 OSes (OSX 10.6-32, OSX 10.6-64, CentOS 5.5). We had to purchase or repurpose Macs (*grumble* stupid Jobs and not allowing VMs */grumble*) for each developer.

*And, one of our new policies is to keep these machines in operation, exactly as is, until we officially drop support for a platform. (This saves us getting into the old LassoSofts problem of not having the recursive ability to re-release an old product.)

Add to this the fact that Lasso 8 is a million internet years old, it has actually grown up through various releases of various OSes, and you have a background set OS support issues worse than Windows printers.

Tin can math on this shows that building and testing for a SINGLE OS with the new team is valued at about $10-15,000. (I still don't have a final figure). Fortunately, we discovered a special ceramic pot of infinite cash which we can dip into at our leisure, and years of time to get Lasso back up to competitive bench strength. Oh wait, no, that's not true.

So, we did a survey to better understand our problem.

What we found, based on survey responses, was this;

SURVEY QUESTION: What is the most important platform for you for the next release of Lasso 8.x?

We are covering 82% of the core sense of urgency with OSX and CentOS. Windows altogether only accounts for 10% at the moment.

This is also exacerbated that developing for Windows, specifically, is a different brain-space than Linux and OSX, which have more similarities (obviously). However, we have also addressed the realities that many customers of the old LassoSoft are still impacted by these issues, and would be willing to pay for a fix - which helps us economically improve future versions of the product.

Based on analyzing the survey results (thanks Rachel), we have also been able to draw conclusions about who is developing on what platform. Here are the results;

Clearly, Windows still plays a key role in our community. More importantly, how many servers are there to deal with;

This information also affects pricing (which I will discuss soon). Should we charge $20k per license, then we will be getting return on our investment for each OS we do a release for which is purchased. I strongly believe the market will not bear this price (hehe - though it has been suggested by an external consultant). The current standing price for Lasso is probably about right. (More to come on this in a later blog). This becomes even more obvious when we look at the facts of what people are willing to pay;

Some tin can math (the realities of economics right now);

Server 2000 has 3 people willing to pay about $600.
Server 2003 has 18 people willing to pay about $4600.
Server 2008 has 21 people willing to pay about $5200.

It doesn't take a business genius to say: "We need to see how a release of 8.6 on OSX fares before we expend resources on a release for Windows". I ain't no business genius, but I can add.

This means, by simple business logic, we should focus on products which have equivalent or logical return on them so as to allow us to grow and improve the business and maintain our competitive advantages of speed, security, simplicity etc. So, logically, we made a decision to develop for the two OSes which clearly are important to people on first pass, thus clearing out the core bubbles;

OSX 10.5-6
CentOS 5.5

After we ensure that there is a market for these two products, we will address returning to any other operating systems for which there is clear and strong economic (and not emotional) demand. We need to be cognizant of the future as well as the past - perhaps more so.

So the actual answers, based on this information, for the following questions;

Q: What is a complete list of operating systems 8.6 will support?
A: For now, see above. After the first release and user demand is clarified, the sky is the limit.

Q: Will 8.6 support Windows?
A: We'd like it to! We plan it to!

Q: What is your timeline for releasing 8.6 on Windows?
A: After we release OSX and CentOS versions, we will address the demand and circle back. There is much to be done, and the Lasso community has spoken. We aren't abandoning, we are prioritizing.

I guess my last question is: when CentOS is so wonderfully stable and easy to work with these days, costs nothing, can be run virtualized for convenience and easily duplicated, doesn't have the security and vulnerability issues of the average Windows server:  and most of all, the Windows world has largely deserted us into the "no one but us" mentality of .net infrastructure: why is anyone still on Windows?!?

*cough* Of course, Treefrog runs LEAP 3.0 (i.e. Lasso 8.x) on various Windows servers for many clients (including GE, for example), as well as hundreds of sites on CentOS 4.7 servers, so I am certainly self-determined to release for other OSes and versions if only to save my own staff some long term hassle. *cough*

SPECIAL NOTE: If a specific individual has a specific task or OS they wish a Lasso build for, and is willing to sponsor the direct full time and materials for another OS specifically, we would be happy to provide the time to do so and prioritize it accordingly. It would then be made public for the rest of the world. Otherwise, it will come back to futurecasting the "next biggest priority" of the community and making sure we address this, possibly before cycling back to Windows. AmigaOS anyone?

Note that I'll post more answers to surveys over the coming week. Watch this space!

Sean Stephens
CEO
LassoSoft Inc.