Revolution - 2.1rc1

I have been developing applications with Revolution for a number of years and participated in beta testing of version 3. To me Revolution 3.0 is a very interesting release for a number of reasons.

Runtime Revolution seems to have streamlined their development cycle. New beta builds were being pushed on a pretty regular basis. As testers this was wonderful as we could test new features, such as multi-dimensional arrays, provide feedback and bug reports and then see fixes shortly thereafter. I've never seem them progress so quickly on a new release.

Fixing bugs seems to be a very high priority for Runtime Revolution right now. Starting with 2.9 and continuing with 3.0 they have been fixing lots of issues. My confidence in the company's commitment to creating a quality product has increased quite a bit.

Version 3.0 added the ability to assign multi-key accelerators to your application menus. Previously you could only assign menu shortcuts that used Command + a single letter/number. Now we can assign multiple modifiers and use keys such as arrows. I don't know that this is one of those features that will sell lots of new copies, but it definitely helps us developers provide more polished products to our customers. I'm glad the engineers are taking time to polish off features such menu accelerator keys.

3.0 introduced the multi-dimensional array to Revolution. This is the first addition of a data structure in all of the years I've been using the product. Having data structures that can represent hierarchal data such as XML is a huge productivity boost for me. I'm looking forward to further enhancements such as this in the future.

Having used this product for 2 years, I don't consider it a viable option for serious software development. It's more for the hobbyist category or rapidly designing an interface for getting things done quickly.

The english-like nature of the language is often presented as an advantage. That's a two sided sword. The problem of english-like and very high level language is that they rely on a huge vocabulary of terms. It's difficult to keep a good mastery of the language when you don't program it on a very regular basis. Another problem is that as a non OO language, there is no easy way to write reusable components. Evidence of this is the lack of components as obvious as numeric stepper, tree, decent table object.

On the lists, the user often comment on how helpful members of the community is. My experience is that many of them are very helpful indeed. It is difficult, however, to have discussions on advanced topics. like in unit testing, project management, design, and software development. When discussions take place on these topics, they are often shallow. It goes with the community as it goes with the language. Your first impression will be very positive, "hey, these guys are cool and helpful as I have never seen on any other iist". However, you will rapidly discover that to get critical information about the software and its features, you have to spend a *lot* of time on the list. The more time you spend within this friendly community, the higher your chances to feel "recognized".... but also the higher your chances to isolate yourself from other persons that are more suitable to hang around with if you have for ambitions to keep up with advances in technology and recommended software development practices.

Going for the "easy" route by adopting a software that let you develop in 2 months an application that that would take up to 6 months in a language like java and c++ can have its short term rewards and can encourage you to believe that you have become a professional software developer. The problem is that there is very little room for progression once you have reached that point. This may perfectly suit your goals, but for the rest of your career, you will be able to develop small-sized to middle-sized applications. You won't be able to move to the next level. If at some point you decide that you need to move to another level, the nature of the language, which is to a large degree unlike any of the common languages, will mean that the baggage/acquired experience you will be able to transpose to your new development tools will be limited.

But the most important problems are perhaps more with the company's attitude than are a bad omen for future improvements of their software. This is a developer tool. That is a tool that developer are supposed to use to provide quality software to their clients and their customers. The company doesn't seem to be committed to serving this category of users the way they could expect to be.

Some declare this the most reliable version ever. The problem is that this qualifier is being used because the product has had a reputation of lack of reliability and there have been small improvements over the last months that get some hope the trend will continue. I am not as optimistic as other reviewers on the fact that it did significantly improve over the last months.

In late 2006, the company toke a stand on quality, promising to improve their processes so to produce better quality software. The first version released after these promises, version 2.8.0 showed important evidence of lack of care. A central feature of the application, the dialog boxes showed in a completely improper manner. Despite mentions of the problem in betas for 2.8.1, the problem was left unfixed. Then, as the user base insisted, it was marked as "fixed" by the director of technology when, well, the code had been modified but the problem hadn't been properly fixed.

Things started to get fixed with more care only *after* I sent an email to their support to warn them that if they didn't provide me with a software meeting minimum requirements of quality, I will consider suing them for breach of service (yep, that's how happy I am with this product). My email didn't get any reply. My account to their bugzilla website got cancelled without any warning. They however extended the beta by another few weeks and made sure they fixed the most obvious problems.

But that a few things got fixed doesn't mean that the software provides with features that are at the level of quality you would expect from a software in that price category. PNG pictures, in particular, behave poorly under image flipping. Image rotation is simply something you cannot rely on.

As an enterprise user, I had been promised in November 2006 (1) 2 support incidents included free each year, (2) Runtime staff now commit to posting weekly on specific and pertinent issues (this isn't me or other non-engineers). Well, some of my emails to support simply never received an answer. The ones that did, where I was promised that something would be taken care of didn't have it taken care of. As of late March 2007, there had been no single post on the enterprise user list by the runrev staff. As I sent an email to make unambiguous that I expected them to honor the promise that had been made, my enterprise user account got canceled, without any warning, and with my emails to support on this problem left unanswered. Rumors had, however, that short after that, their staff take a bit of time to answer user questions and that their CEO posted an informative email on the list. I have not heard of any rumors that this continued on a regular basis.

In the advertising emails that you sometimes received after downloading a trial software, it is stated that the application let you build native applications for Linux. This is not true. Support for the build of Linux application stopped, with version 2.7, released in January 2006 and hasn't been resumed since then. If you ask, they will probably tell you that this will be made available very soon. Sure, they have presented the release of the Linux build version as "imminent" for a year and a half. To the best of my knowledge, as of today, nobody as seen even an alpha version of this "imminent" Linux version and no release date has yet been announced.

On the "potential" of the software, I would rate it as excellent. On the real thing, including the absence of extended components, the poor support, and the poor attitude of the company, the scarity of third party development (and the total absence of incentive for a could be third party developer given the very small size of the community), I would rate it as "very poor". I rate features as neutral because though they recently added the ability to include a browser or sqlite database within their app, these features were originally developed as plugins by an independent developer. In contrast, basic features like usable image rotation, basic components like numeric steppers, decent table object, graphing library are missing altogether.

I would recommend to wait until *after* the promises they made of an astonishing 2.9 and then 3.0 have been honored, to consider buying the product.