DesktopLinux
Home  |  News  |  Articles  |  Forum  |  Polls  |  Blogs  |  Videos  |  Resource Library

Keywords: Match:
DesktopLinux.com interviews Gnumeric project leader Jody Goldberg
Rick Lehrbaum (Jan. 11, 2002)

After three and a half years of development, the Gnumeric project recently announced the availability of version 1.0.0 of Gnumeric, an open source spreadsheet program. To learn more about the what goes on behind the scenes in an open source project and what makes an open source developer "tick", DesktopLinux.com founder and executive editor Rick Lehrbaum chatted with Gnumeric project leader Jody Goldberg. Here's what Lehrbaum learned . . .



RL: Briefly describe the purpose and of Gnumeric and its current status.

Goldberg: Gnumeric is a spreadsheet. The goal is to develop an application that will let us get our work done. This means offering good compatibility with existing applications and providing a solid architecture so that people can extend the project to suit their needs. After 3.5 years of development, and a lot of testing, we have now branded an officially stable release as 1.0.

RL: What is your role on this project?

Goldberg: I'm the primary maintainer. Most of my time is spent on basic janitorial duties, such as bug triage, and code consistency. At other times I review patches, extending the system as a whole, and make useful pronouncements from on high. Such as: "The docs shall call the key 'Enter', not 'Return'"

RL: How long have you been in that role, and how long have you worked on the project overall?

Goldberg: Miguel de Icaza created the project during the summer of 1998. I began contributing spring of 1999 and started making the releases in the fall of 2000.

RL: How did you come to be working on this project?

Goldberg: My day job was in finance, working for various investment firms in New York, and spreadsheets are an ideal way of deploying and testing new analytics. After several years of frustration with the limitations of various proprietary products I decided to contribute to a solution. A program like MS Excel (tm) is tremendously powerful. Microsoft has worked long and hard to produce a very functional tool. As long as it works, and as long as you are working the way it expects, you are fine. The moment you want to do something they have not thought of . . . boom. You are out of luck. After wasting days fighting my way up the "tech support" ladder to find someone willing to admit that I was seeing a bug it became clear that the only viable solution would be to have source code.

RL: How much of your time do you typically spend working on gnumeric?

Goldberg: As much as possible. The majority of the time is late nights and weekends. However, my 6 month old son (shown in the photo along with Jody) ensures that there are lots of late nights to go around :-)

RL: Please describe yourself for our readers (as much as you care to), including your personal/professional background, occupation, hobbies, likes/dislikes, interest in linux, etc.

Goldberg: I'm a very lucky guy. When not working a project I attempt to be a dad and spend time with my wife and son, and our dog.

RL: What is your philosophy/attitude with regards to open source software in general, and linux in specific?

Goldberg: In my opinion, developing in an open environment enforces much higher standards on the quality of the code developed, compared to proprietary applications. Look at the amount of code review taking place. However, it also removes the direct accountability that forces people to work on more mundane aspects of a project. A fair amount of discipline is required to ensure that the core of a project is solid, and you are not just adding "neat" features.

Linus' recent comments on "directed evolution" are absolutely correct. The software engineering concept that you can know, up front, what the real requirements are does not hold true in practice. You rarely know the right questions to ask before starting a project, and you certainly do not know enough to design a good data model. The only viable solution is to keep the core clean, and to replace things as the problems present themselves.

RL: Is any company directly "sponsoring" the gnumeric project? If so, who, why, and to what degree are they providing funding and/or programmer time?

Goldberg: Miguel de Icaza began Gnumeric before International GNOME Support was created (which became Helix Code, which became Ximian). Without Ximian's support the project would certainly not be where it is. While doing the credits for the 1.0 release announcement it became clear that a significant percentage of Ximian employees had directly contributed to Gnumeric at some point. Although it is not a core Ximian project, it is certainly a well connected mascot.

RL: Is the gnumeric project associated with a particular group of open source projects, or is it basically an isolated project?

Goldberg: Gnumeric is part of the GNOME project. The spreadsheet aspect is basicly stand-alone, but the application framework is layered on top of many GNOME tools. As a large application we were frequently used to test new technologies, such as gnome-canvas, gnome-print, and libglade. As those matured our contributions shifted to bug tracking and leak plugging. We would get the bug report when a library was in trouble. Thanks to Morten Welinder and his diligent application of "Purify" Gnumeric and all the libraries we depend on are leak free (we hope).

RL: Is gnumeric usable with KDE, or just GNOME?

Goldberg: Absolutely, we've had several compatibility requests mainly dealing
with 'Klipper'. Gnumeric uses a full MVC separation, which keeps the gui aspects at arms length. I've started looking into offering an xparts wrapper once development moves to gtk-2.

RL: Is there anything else you'd like to say to the Desktop Linux community?

Goldberg: Two thoughts . . .

First: This is not a race. There are frequently posts and articles asking when linux will be ready for the desktop, complaining that it is not ready for the desktop, or pontificating that it never will be. This is nonsense. It has taken years of development for MS and Apple to produce their frameworks. They have more money and more beta testers (aka paying customers). Open source development is not magic and can not instantly overtake these projects. The desktop is an evolving target, we are getting there quickly. There are already many users whose needs are met by open alternatives. That number will continue to grow. That is the beauty of open source, it can continue to improve and expand as long as people are interested in it.

Second: There is not enough Innovation. Beyond the pollution of that word, there is the question of how to balance improvement with compatibility. There are those at both ends of the spectrum. People who advocate slavish replication of Microsoft's choices (be they good or bad) and those who would prefer to see radical changes made. IMHO neither extreme is a good idea. If we are compatible enough that people can move back and forth smoothly the user will be happy. If we make applications that are easier to use, people will also be happy. Microsoft is the master of this strategy. It is called "Embrace and Extend", and I for one believe that turn about is fair play.

RL: When you said "There is not enough Innovation", did you mean that the linux community isn't innovating enough, or did you mean that those who attack linux as not being sufficiently innovative are not right to do so?

Goldberg: The latter. There are people who make that complaint (that linux isn't sufficiently innovative) and I disagree with them. We've had criticism that we are not using gnu-step, or emulating Quatro, or using mathematica's interface. Less specific comments complain that Gnumeric/Ximian/GNOME are just cloning Microsoft products. In my opinion, these miss the point entirely. Office represents the important monopoly. With it, Microsoft controls your data! Any products hoping to survive in that area need to be able to interoperate extremely well. With that compatibility comes a degree of similarity. People want to be able to make the same types of changes with the same effects in your program. We have already started making extensions, but it is difficult to branch off into new territory unless you can support a critical mass of the existing feature set.

RL: It's interesting that you used the expression "Embrace and Extend" just now. Given that the recently released Gnumeric version 1.0.0 had the codename "embrace", that leads me to think that the codename "embrace" meant that Gnumeric would be highly Excel-compatible. Extrapolating from there, it would seem likely that a future version (Version 2.0?) of Gnumeric might be codenamed "extend", and would have the goal of going beyond the functionality of Excel. Care to comment on this conjecture?

Goldberg: I'm working towards that, yes. However, realisticly speaking, Excel has a fairly large headstart in most areas. Guppi is targeted at what we perceive as their weakest link. Graphs are the first major area that can be improved. However, I'd like to get 1.1 tree branched (this weekend) and commit the gnome2 port before we think about a 2.0.

RL: Thanks very much for taking the time to speak with the Desktop Linux community and thanks, also, for your tireless efforts in the very important Gnumeric project.



Related stories: Talk back! Do you have comments or questions on this story? talkback here


(Click here for further information)



Home  |  News  |  Articles  |  Forum  |  Polls  |  About  |  Contact
 

Ziff Davis Enterprise Home | Contact Us | Advertise | Link to Us | Reprints | Magazine Subscriptions | Newsletters
Tech RSS Feeds | ROI Calculators | Tech Podcasts | Tech Video | VARs | Channel News

Baseline | Careers | Channel Insider | CIO Insight | DesktopLinux | DeviceForge | DevSource | eSeminars |
eWEEK | Enterprise Network Security | LinuxDevices | Linux Watch | Microsoft Watch | Mid-market | Networking | PDF Zone |
Publish | Security IT Hub | Strategic Partner | Web Buyer's Guide | Windows for Devices

Developer Shed | Dev Shed | ASP Free | Dev Articles | Dev Hardware | SEO Chat | Tutorialized | Scripts |
Code Walkers | Web Hosters | Dev Mechanic | Dev Archives | igrep

Use of this site is governed by our Terms of Service and Privacy Policy. Except where otherwise specified, the contents of this site are copyright © 1999-2011 Ziff Davis Enterprise Holdings Inc. All Rights Reserved. Reproduction in whole or in part in any form or medium without express written permission of Ziff Davis Enterprise is prohibited. Linux is a registered trademark of Linus Torvalds. All other marks are the property of their respective owners.