Solprovider Thoughts Kongregate Registration for Free Flash Games and Chat with solprovider

IBM screws Lotus


Originally written 20030305 for ZDNet

IBM does not understand that software can be a business. All of their software products exist to support their hardware and consulting sales. DB2 drives AS400s. (Does anybody ever put the "?-series" name without putting the old name next to it?) Websphere is an attack on Sun. Tivoli is a tool for the consulting division.

And then there is Lotus Notes. The program with a following of non-IT people that won't die because it is more useful to business than any other software since the spreadsheet (which was also originally made popular under the Lotus brand.) Lotus Notes developers are some of the highest paid IT people today, they have less training than VB programmers, and they are each more productive than medium-sized development teams.

When a Fortune 100 company needs to replace their mainframe contract management system with a new application designed, coded, tested and in production in less than a month, they use Notes. And the entire project is handled by ONE developer.

When an amusement park needs a web portal for use by all the employees, they hire ONE Domino developer for a four month contract. And he leaves after three months because they cannot think of any more features to add.

I fell in love with C because it was more my style than Pascal or BASIC. I liked C++ even more because it allowed me to blackbox code. I love Java because it adds the ability to easily multi-thread, along with incredible error catching.

I use Lotus Notes because it allows me to build incredibly complex, secure, useful, scalable and maintainable applications in less time than it takes to document the specifications.

But this is about the current article.
Rather than pump up the world's best software with all the abilities of its other software, IBM will dismantle it and hope that some of the charm rubs off.

"Maybe if we move all the applications to Websphere then people will buy WebSphere rather than Notes."

WRONG!!

Businesses use the applications because they want Notes and using the stock or add-on applications is easier than building their own. But unless IBM seriously hurts the ease-of-use and easy-development in Notes, it will be much, much easier to develop an in-house application on Domino than to configure WebSphere.

"Lotus' day has finally come!"
First IBM bought them.
Then the original Lotus management moved on.
Then IBM forced Lotus to not compete with IBM's other products.
Then IBM chops up the Lotus Notes applications and hopes the pieces are worth more than the whole, not caring if they are not.

The new leader says "that groupware and collaboration software hasn't lived up to its earlier promise."
Has he used Lotus Notes? Has he ever used a Notes application? Has he used one written by experienced non-IBM developers? Has he asked his secretary to build software using Lotus Notes? What promise did Lotus Notes fail to deliver? Teamrooms can be set up in minutes to allow global asynchronous communication.

The biggest problem with most Notes networks is that applications are so easy to create, they ARE USED. But projects can last a week (such as writing a proposal) to years (such as the corporate portal.) The problem is tracking which applications are dead and just wasting hard drive space; they have served their purpose and should be removed.

This is considerably different than other platforms where the main issue is to have a working application before the project has ended. Employees will never ask for an application to help write a proposal if the proposal is due this Friday and the minimum development time is 2 months.

IBM hopes Notes will die if they shove DB2 up its backend, steal all the applications for WebSphere, and remove all its funding. Any of these choices will severely hurt Notes. Together they may actually achieve the goal.

"Lotus' day has finally come!"
Say goodbye.


Reply by Roland Weber 08:35 GMT 03/05/2003

Everyone at IBM has used Lotus Notes, because it's IBM's internal mail system. And Sametime is quite popular as well within IBM.

But customers have loads of HTML-based applications. And they will not port them over to Lotus. But they are *very* likely to integrate them in a portal, such as the WebSphere Portal (WPS). Only problem is - Lotus applications do not yet integrate very well with portals. So the customer has own applications, has WPS, but won't buy Lotus Notes or Sametime because it doesn't integrate well enough.

As for building scalable applications on Lotus, I wonder what you refer to as "scalable". Lotus sure scales well to hundreds and thousands of users. But WebSphere in general and WPS in particular is meant to scale to millions of users. You won't do that in three months time without finding more features to add. Even if it's only for the administration and management of the application.

Disclaimer: I'm an IBM employee and my work is centered around WPS.


My reply at 18:20 GMT 03/05/2003

I know everybody at IBM uses Notes. I have code running there. (But I have never been an IBM employee.) My phrasing was deliberate. IBM Notes applications have to meet so many standards that the development effort suffers greatly. The ubiquitous use of Navigators make the interfaces pretty, and the inability for an Editor to add Views prevents some performance issues, but they hurt the ability to use Notes to the fullest.

As far as scalable, I also have Notes applications running at GM. Scalability is a major issue for any development there. But the ability to put applications on many servers around the world makes it usable. And Replication keeps the data current.

I agree that WebSphere should be integrated into Domino. IBM has done a great job on the HTTP stack and the servlet manager in WebSphere. These should be integrated into Domino. Domino can integrate other systems easily.

Integrating Domino into WebSphere may be impossible due to the different paradigms. Domino provides a fantastic security model. Domino provides a great front-end. As a web server, it provides rapid application development, easy maintenance, and total integration with other systems. Some of that integration is because WebSphere and LEI can become just another part of Notes. But Domino cannot become "just another part" of WebSphere without losing all the abilities that make it great.

Most Notes applications, even those written by non-IT people, can be used productively the same day they are written. Yes, they are unpolished, and it is quite likely that much functionality will be added later, but they are usable.

Most companies have a standard teamroom template. It is overkill for most of the times it is implemented, but it works and can be deployed in 10 minutes by a non-IT person (except at IBM.) Referring back to my examples, a manager needs an application to work on a proposal. She creates a teamroom database on a server, and her team is productive. They win the proposal, and use the teamroom to write the specifications. They do the job, discussing the work in the teamroom. The IT group is never involved. (And the database is still there 10 years later because nobody knows it is obsolete.)

Your belief that useful applications cannot be written in 3 months shows that you do not understand the Notes paradigm. Small appplications delivered quickly means productivity.


Monday morning: Meet with business manager and an employee to discover needs.
Monday afternoon: Create new database. Add Forms. Add Fields to Forms. Add basic Views.
Tuesday: Add functionality. Make it pretty.
Wednesday: Meet to display prototype and create alternate Views based on use of data. Verify security model.
Thursday morning: Finish functionality.
Thursday afternoon: User acceptance testing. Fix anything found.
Friday: Delete test data and announce availability of the application.
(Integration with a backend RDBMS may add a second week of meetings.)

At least half of the time is dedicated to working with the business people. And this is for a CUSTOM project. The code for this application can be redeployed in 10 minutes by a business person if the business practices allow for it (as in the teamroom example above.)

I understand this is a small project. But most projects will be this size if you allow the business people to take full advantage of technology. Otherwise they spend their time emailing spreadsheets to each other.

There is a place in the world for the corporate-wide projects, such as payroll or CRM. This is not where Notes is the most useful, although it can help by being the front-end to the either of these systems.

WebSphere and the other "application servers" have a place in the world. Many applications will be written on them. But they are limited to those applications that require massive investment before productivity. And that suffices until people learn the possibilities of Notes. Because Notes will help with today's tasks, not next year's. And people like that.


Reply by Roland Weber at 06:25 GMT 03/06/2003

You wrote:
"Your belief that useful applications cannot be written in 3 months shows that you do not understand the Notes paradigm. Small appplications delivered quickly means productivity."

I never intended to say that *useful* applications cannot be written in three months. My apologies if I didn't make that point clear. What I intended to say is that large-scale applications are not written in three months. And WPS is a product targetting large-scale applications.

The integration of Lotus products with WPS is meant to fit Lotus applications better into the large-scale portal world. And your original posting stated that this was a bad idea. Or at least I interpreted the subject "Doing it the wrong way for all the wrong reasons" that way. I just wanted to point out some reasons why it is not the wrong way. But I'm *not* saying it is the only way, or that the current line of Notes/Domino products is a bad idea or whatever. Both have their place in the IT world. Grant me my integrated WPS world, and I'll grant you the world of quickly deployed and immediately useful small applications :-)


My reply at Thu, 06 Mar 2003 21:07:04

I agree that WebSphere has a place TODAY. But if IBM developed their software properly, that market would disappear quickly. With IBM's animosity, it may take well over 30 years. If IBM would help, it could take much less than 15 years. IBM is the crucial company because they own Lotus Notes, which contains the kernel of the computing algorithm of the future. If IBM ruins it, then I will probably spend the rest of the decade developing its replacement.

I am living my beliefs. We are developing a "portal" product that will plug the holes in Domino. It should allow Domino to compete with the "large-scale portal world," although the product also takes portals to a new level. We expect to start a corporate beta test next quarter, with a sales push following soon after. It complements WebSphere, since it requires a Java Servlet Manager to be added to Domino 5, and the only choices are WebSphere and Tomcat.

You can target "large-scale applications". And I will be able to run them on my wrist-watch in 20 years. Hardware advances will remove the requirements for "large" computers. Software development needs to advance too. I will push the advances in software develoment if IBM refuses. (I am not saying I will be successful, but I will try.)


(Not Posted)

You and most of the world believe in the top-down approach to corporate systems. I believe in a bottom-up approach. As in software design, you need to mix them.

For example:
Normalization suggests there should be a single record for each employee. But each department will want different data on that record. Human Resources needs the skill set. Payroll needs the current pay rate and deductions, time worked, and where to send the money. Managers need time worked on each project, goals achieved, and status of current projects, as well as the ability to enter a raise. IT Managers may add known programming languages and platforms. Marketing Managers may add known written languages and knowledge of cultures. A single system could hold all this. I do not yet know how to make it both efficient and fully normalized, but I can make it work.

I see 3 types of programs:
1. Solo. Use by yourself. (Most games, statistical analysis tools, compilers.)
2. Piped. Single connection. Classic client-server where today's hardware requirements keep it out of the third category. E-commerce.
3. Collaborative. Used by a team or a company or even many companies. (Groupware, forums, extranets, internet gaming.)

Today we can see the second category blur into the third. Check out Amazon.com. Here is our product list. Here is an order form. But wait, there's more. You also get recommendations from other users. You can also share your wishlist with your friends. Collaborate!

Piped programs were necessary for performance reasons during the early days. Hardware improvements have made it possible for all Piped programs to move into the Collaborative realm. Even many Solo programs are becoming Collaborative.

<< LN and MSAccessUbiquitous Lotus Notes >>

Contact Solprovider
Paul Ercolino