White Paper: CM: THE NEXT GENERATION - Where Are We ... Going?

CM is picking up the pace. After the longest time as a Version Control solution, CM is now commonly viewed as a Change Management solution. The difference is not primarily one of functionality. It's a difference between designer utility and product backbone. As such, where Change Management has been embraced, CM is now center stage. And that means it will be a catalyst for full ALM solutions and eventually enterprise business management. What are the key achievements that have happened recently and what advances are needed to embrace a wider solution space?

A transition is happening from one extreme end of the CM solution space, to another. It may take a while to complete. With some products you may need to cut the definitions a bit of slack (e.g. what is "end-to-end") but with others, it's the market demand that needs to catch up to the offering. Here are the two extremes:

A: Complex, Rigid, Big-IT "solutions"

B: Lightweight, Flexible Solutions

Why do I think this transition is gathering steam? Well the primary reason is that the industry, that is, the user community, is no longer tolerating the status quo. Even though we have quad CPU processors, GBytes of memory and high bandwidth network capabilities, it turns out that the small footprint solutions are easier to deploy, generally more scalable, and tend to incorporate newer, more advanced technology. The fear of not having enough, if it's not a Big-IT solution, has turned into a fear of not having enough if it is one! And rightly so in many cases.

Big-IT means lots of administration of resources, complex admin processes, loads of training. When I look at a CM tool, I want to install it, load in a representative chunk of my project and get going with it right away, in hours from first consideration. I don't want to have to buy additional hardware, bring in consultants and go through a bunch of training, just to do an evaluation. It's fine to take time to put it through all of the paces and check out its scalability, functionality and reliability - but not so fine if I have to put in an equivalent amount of effort up front, and even less so if I find that after doing so, it doesn't measure up.

And as more capabilities come on line, Big-IT, glue-integrated tool solutions don't make the grade. Consider Multiple Site operation. Not only do you have to have a CM plan for this, but one for all of the other tools in the solution. Then the synchronization operations must be coordinated. Another question that comes up is "How do you do consistent backups?", especially when shops are operating 24/7 around the globe for your enterprise! What about process improvement? If I have to learn 4 process tools to implement process across my lifecycle, forget it.

So as "small-IT" solutions gain credibility, more and more organizations are looking at them. If they fail to cover the end-to-end spectrum, they too will fail to make the grade. But there are "small-IT" solutions that are starting to measure up to the full ALM problem, some slowly and some more advanced. And that's the feedback we see. Sure we still get the occasional "Technically your solution wins hands down, but you're not big and blue...". And that's fine, because it's the trend that's important. The RFI input we see includes, more and more, things like a need for:

The industry is starting to realize that it can ask for it's cake and eat it too. And so the transition continues.

Where are we
So where are we now? Well, most CM tools are still 2G with a few 3G features. But the tools are advancing in technology more quickly. Especially the newer tools and the ones that have a single ALM architecture. Ease-of-use is important today, and this rules out a few of the older solutions. We're in a race. New tools are trying to expand their functionality while older tools are trying to become easier to use while addressing a wider ALM solution. As a start, almost everyone is demanding:

And with easier to use CM, it's only natural that we try to do more advanced CM. In our own shop, we like to stay ahead of the curve where possible. We've started using more and more features over the past few years:

Where are we... going
But the story is just starting to get good. There are a number of innovations coming foward in the industry over the next year or two.

The Virtual File System monopoly which has been dominated by a single vendor product (i.e. ClearCase) over the past 15 or so years, will give way to one or two other vendors, with a focus on less resource intensive operation. What this means is that we'll see more tools offering access to both static and dynamic context views of their configurations, directly from the operating system. If the traditional server bottlenecks are eliminated for this type of technology, and the associated administration eliminated, we could see a wider market beginning to adopt CM functionality (e.g. documentation groups, accountants, etc.)

Multiple Site operation will approach zero administration operation as automated synchronization and monitoring capabilities enter the picture. Data segregation will still be important, but this will be a logical specification that will be supported with a simpler operating framework that does not require data partitioning and synchronization management skills.

At least one CM vendor will be introducing a solution which caters to the CM II process (see ICM). Expect to see others that cater to the CMMI process. Ideally, these will be optional and configurable starting points for a solution rather than rigid configurations.

Customizable dashboards for product, project and verification status will begin to take center stage as the tools become more and more management oriented. The value of CM tools as a backbone product management tool, and as a backbone for business management, will be exposed and the rush will be on to include management more easily in the default, and even driving, set of users. Dashboards and central data repositories will make CM tools central communication hubs - so much so that users will prefer pulling information from the CM tool rather than having a myriad of emails sent to them through customization of triggers.

Broader and easier customization of the user interface, and the CM process, will take over from the more rigid fixed process configurations. There will always be room for tool personnel who can work wonders with the product. But more and more of the basic customization capabilities will be simple menu-driven operations: adding a new data field or a new traceability link, specifying/evolving state-transition diagrams, establishing better security, etc.

Better out-of-the-box, end-to-end solutions will appear with rapid traceabiltiy navigation. The days of glue departments will begin to disappear. As well, end-to-end solutions will be seamlessly integrated, through a single user interface running on top of a single repository, as opposed to one per application area. Operations such as backups, multiple site operation, etc. will work across all applications, not just the file-management areas.

Ever wider application suites, embracing things like customer management, time sheet tracking/project roll-up, project management and full traceability between Requirements and Test Cases. Not only from a coverage perspective, but also from active test run data against specific builds and configurations.

And CM Vendors will begin to offer free on-line or webcast courses to push their solution through to the short-lists of prospective customers. When you see this happening, the market will have turned a significant corner. Ease-of-use for CM will have matured. CM training and consulting will focus on process, not on tool usage. And the tool vendor will, in most cases, no longer be the main provider of CM training.

When will these things take place? Well, they've already begun, and will continue this year and into the next. And as each vendor contributes to the advances, the others will be forced to play rapid catch-up or to drop out of the race. The market expectations will move from one of tollerating CM tool behaviour to one of a backbone team support role. The pace will continue to quicken and a few vendors will emerge from the myriad now out there, as they mature in their 3rd and 4th generation solutions.

Open Source
Will the open source community eventually take over? This is a hard call. Compilers, GUIs, Operating Systems, etc. have fairly well-defined requirements and if you can use one OS, you can likely use another. CM is complex, very complex. Not just it's nature, but also in its diversity. Go from one site to another and you have a new set of requirements: hundreds of little products, legacy languages, thousands of team members, different IDEs, specific contracts vs. general market, end-user product vs. system component, etc.

The customer for a CM system varies a lot more than for the other Open Source products. So while open source will continue to produce better point tools, process and customization requirements are sufficiently diverse and complex that Open Source tools will either not be able to embrace the full set sufficiently, or will do so in too complex a manner. Look at how difficult it's been to get a good desktop configuration of Linux for example - one that a significant market share could embrace. I'm afraid the task is more difficult for the CM world where the user community is much more widely dispersed.

But then there's always the wild card - will one of the vendors turn over it's assets to the open source community - especially if they find it difficult going alone in the upcoming race? Well, not this year anyway. Open source will focus more on knitting together point solutions to provide a basic ALM solution. But the long-term viability of this approach will be uncertain, at best, because of the added complexity that comes from combining such solutions.

Joe Farah is the President and CEO of Neuma Technology . Prior to co-founding Neuma in 1990 and directing the development of CM+, Joe was Director of Software Architecture and Technology at Mitel, and in the 1970s a Development Manager at Nortel (Bell-Northern Research) where he developed the Program Library System (PLS) still heavily in use by Nortel's largest projects. A software developer since the late 1960s, Joe holds a B.A.Sc. degree in Engineering Science from the University of Toronto. You can contact Joe by email at farah@neuma.com

More white papers:

Find out more ...

Neuma White Paper:

CM: THE NEXT GENERATION - Where Are We ... Going?