The Xxl Spreadsheet Project
During the time frame of my academic work, I must periodically deliver statements of student grades to the administrative offices. The number of grades I manage is not immense, about two thousand per year, but it is enough to warrant automating the computations with the help of a spreadsheet program. My daily computer environment is Linux and until 1996, I was regretting the lack of a public domain graphical spreadsheet program for the X Window System, one which was user friendly and simple to use.
Of course, spreadsheets have been available in the UNIX world for many years, but in my opinion, none of them met the criteria of freeness, simplicity and user-friendliness I desired. I was forced to use a commercial spreadsheet on another computer with a different operating system in another office. This situation was not convenient, so I decided to build my own spreadsheet. At the same time, there was a call for proposals of computer projects in the Computer Science Master's program. This was a good opportunity to launch the project and I submitted this idea. It was chosen by a team of four students, thus, Xxl history began.
The origin of spreadsheets is in the world of accounting. This name refers to paper sheets used for computing cost differences between offer and demand, and more generally, between two prices. These computations were made by hand and were tedious and error-prone. The first theoretical work on computer versions occurred in the early 1960s. The main principles of using matrices, cells and budget simulation, as highlighted by these theoretical works, were implemented in FORTRAN programs for simulating enterprise budgets (see Resources).
However, it was not until 1978 that two MIT students, D. Bricklin and B. Frankston, designed a computer spreadsheet, VisiCalc, that was truly usable on a personal computer. VisiCalc allowed many small enterprises to use costly computer techniques previously affordable only to larger companies. This program contributed significantly to the success of personal computers, especially the Apple II, for which the first version was written. VisiCalc was a tremendous success, but unfortunately for its authors, a very short one. At the beginning of the 1980s, VisiCalc encountered difficulties in trying to deal with the rise of IBM PCs. It was replaced by a new spreadsheet based on VisiCalc, Lotus 1-2-3 by M. Kapor. New features of Lotus 1-2-3, such as a simpler denotation for cells, the concepts of cell row or macro and the addition of graphic handling quickly made it a market success and a de facto standard.
In the late 80s, after Macintosh opened the door, most operating systems began to provide graphical user interfaces. The spreadsheets Quattro Pro and Excel appeared at that time and made sheet handling simpler, thanks to the use of tear-off menus, buttons and dialog boxes. Today, Excel is one of the most widely used commercial spreadsheets.
For many years, the UNIX world did not seem to be interested in spreadsheets. The first free and open spreadsheets available were sc and ss (an sc variant). They more or less offer all the basic functionality expected from a spreadsheet, but have no graphical interface. In the early 90s, Xspread (based on sc) and Oleo from the GNU project were the first spreadsheets with an X interface. Both provide new features (Xspread can generate graphics), but their graphical interface is still awkward and lacks user friendliness (none of them offers tear-off menus or provides for the X-style cut and paste).
Today, thanks to the qualities of its kernel and applications, the success of Linux is such that this system is beginning to overshadow some great builders of commercial operating systems. Linux is beginning to enter the industrial world—this is one of the reasons for the present burgeoning development of office suites (StarOffice, Koffice, Applixware, Gnome project's Gnumeric) which aim to equaling their commercial rivals.
Xxl has been designed for ease of use with a user friendly interface. It does not try to compete with spreadsheets such as Excel or Lotus 1-2-3, but will handle small and medium-size sheets.
Most often, due to lack of time or competence, the programming projects of the Master's degree in Computer Science of the University of Nice—Sophia Antipolis are either not completed, or the final result cannot be used in practice. Expecting a usable result should not be a priority, since the goal of the project is mainly pedagogic. For this particular case, I wanted to get a usable preliminary version of the spreadsheet, even though I anticipated some parts of the software product would need revision, as was the case. Thus, the goal of the project was clearly defined and the spreadsheet was limited to the basic functions of any classic spreadsheet, with a flexible and user friendly interface. The project was divided into two different parts, as little connected as possible. Two students dealt with the spreadsheet kernel and two others worked on the graphical interface.
Since the spreadsheet functionality was limited from the beginning, it had to be written to allow for subsequent extensions; therefore, choosing the programming language was of utmost importance. I wanted a language that would facilitate writing a spreadsheet prototype and enable an incremental development method. It also needed to be easy to alternate quickly between test cycles and corrections. Moreover, it should integrate an easy-to-use graphical library to free the programmer from all cumbersome aspects of X programming.
I chose STk, developed by E. Gallesio, which is an implementation of the programming language Scheme together with the Tk graphical toolkit (see Resources). In fact, it is similar to the Tcl/Tk system, with the Tcl interpreter replaced by a Scheme interpreter. More specifically, STk provides all the power and ease of use of the Tk toolkit from within the Scheme world. Moreover, Lisp (Scheme is a Lisp dialect) has already demonstrated its qualities for software extension (e.g., GNU Emacs). Finally, STk offers an object layer, STklos, which provides for a posteriori reutilization and extensions.
Tk (and thus STk) does not provide in its distribution any specialized widget for representing the computing sheet of a spreadsheet program and its predefined components do not make such a widget easy to build. The students could not devote more than half a day per week to the project. Asking them to program such a widget was out of the question—the project would have been an immediate failure. Thus, it was necessary to reuse some already-built components in order to make the programming task as small as possible.
One of the strengths of Linux is that it offers its users an open world. Thousands of programmers around the world are designing and developing programs, often of superior quality, which they offer for free to the international community. One of these programmers, J. Hobbs, is the present maintainer of a Tk widget called tkTable. Not only does this widget specialize in the representation of computing sheets, but it also had a property that made it the ideal interface between my two teams of programming students, i.e., the spreadsheet kernel and its graphical interface. The tkTable widget provides for associating a data structure (representing the cell values) with the graphic computing sheet. After any change in content of the computing sheet, the widget automatically updates the data structure. In addition, and even more importantly, after any change in the data structure, the widget updates the computing sheet. Thanks to this property, the two student teams could work fully independent of each other.
The second component we re-used was the LaTeX environment, available on any UNIX platform. Xxl uses LaTeX for printing computing sheets and previewing them on the screen.
In June 1996 at the project's end, the students delivered the first usable version and we decided to distribute it freely. To give access to the source text of a software product is to accept the judgment of those who will scrutinize it. I used part of my vacation time to revise the code before the first distribution, which occurred in September.
The next year, to my disappointment, no student team chose to pursue the development of Xxl. Its present state was satisfactory for my own needs and I had no time to devote to its further development. Then I received mail from a new web site aimed at promoting scientific applications for Linux (SAL). They offered to give me a page for Xxl. This site gave the spreadsheet a larger audience, which encouraged me to improve it. Once more, I spent part of my summer vacation modifying the Xxl code, which I signed with a mock name. This time I rewrote almost all the code, with considerable simplifications, thanks to the use of the Flex and Bison parser generators. I also corrected some errors and offered new functionality. In September 1997, a new version was delivered.
At the beginning of June 1998, E. Gallesio delivered a new version of STk, with a small error which prevented the spreadsheet from working correctly. Since I was not using this version, I did not have a problem, but I got many messages asking me to correct it from users who thought the error came from Xxl. I had not been aware that Xxl had so many users. I devoted part of July to the distribution of a working version.
While it is clear that in an open environment like Linux, re-using already built software components is a good thing and should be encouraged, this does come with problems. In particular, programmers who use external components which they do not fully master (even with the source code at their disposal) are very dependent on the evolution of these components. Software tools like Flex, Bison or LaTeX, which have been stable for many years, did not cause any problem. This was not the case for the tkTable widget or for STk, which is still under development and constantly changing.
With each new version of tkTable, it is necessary to adapt it to STk. This represents about a hundred changes in the source code of the widget. Obviously, Xxl must follow the evolution of the tkTable widget, since the improvements are a benefit to it.
The evolution of STk could also be a source of problems. Fortunately, the author of this language is a friend whose office is not far from mine. This made it possible to solve many problems. While Xxl certainly benefitted from the improvements to STk, STk also benefitted from Xxl. In fact, the sheer size of the spreadsheet makes it a good test program, exposing some errors in the interpreter.
|Designing Electronics with Linux||May 22, 2013|
|Dynamic DNS—an Object Lesson in Problem Solving||May 21, 2013|
|Using Salt Stack and Vagrant for Drupal Development||May 20, 2013|
|Making Linux and Android Get Along (It's Not as Hard as It Sounds)||May 16, 2013|
|Drupal Is a Framework: Why Everyone Needs to Understand This||May 15, 2013|
|Home, My Backup Data Center||May 13, 2013|
- Designing Electronics with Linux
- New Products
- Linux Systems Administrator
- Senior Perl Developer
- Technical Support Rep
- UX Designer
- Making Linux and Android Get Along (It's Not as Hard as It Sounds)
- Dynamic DNS—an Object Lesson in Problem Solving
- Using Salt Stack and Vagrant for Drupal Development
- Reply to comment | Linux Journal
8 hours 18 min ago
- Dynamic DNS
8 hours 52 min ago
- Reply to comment | Linux Journal
9 hours 50 min ago
- Reply to comment | Linux Journal
10 hours 40 min ago
- Not free anymore
14 hours 42 min ago
18 hours 30 min ago
- Reply to comment | Linux Journal
18 hours 38 min ago
- Understanding the Linux Kernel
20 hours 52 min ago
23 hours 22 min ago
- Kernel Problem
1 day 9 hours ago
Enter to Win an Adafruit Pi Cobbler Breakout Kit for Raspberry Pi
It's Raspberry Pi month at Linux Journal. Each week in May, Adafruit will be giving away a Pi-related prize to a lucky, randomly drawn LJ reader. Winners will be announced weekly.
Fill out the fields below to enter to win this week's prize-- a Pi Cobbler Breakout Kit for Raspberry Pi.
Congratulations to our winners so far:
- 5-8-13, Pi Starter Pack: Jack Davis
- 5-15-13, Pi Model B 512MB RAM: Patrick Dunn
- 5-21-13, Prototyping Pi Plate Kit: Philip Kirby
- Next winner announced on 5-27-13!
Free Webinar: Hadoop
How to Build an Optimal Hadoop Cluster to Store and Maintain Unlimited Amounts of Data Using Microservers
Realizing the promise of Apache® Hadoop® requires the effective deployment of compute, memory, storage and networking to achieve optimal results. With its flexibility and multitude of options, it is easy to over or under provision the server infrastructure, resulting in poor performance and high TCO. Join us for an in depth, technical discussion with industry experts from leading Hadoop and server companies who will provide insights into the key considerations for designing and deploying an optimal Hadoop cluster.
Some of key questions to be discussed are:
- What is the “typical” Hadoop cluster and what should be installed on the different machine types?
- Why should you consider the typical workload patterns when making your hardware decisions?
- Are all microservers created equal for Hadoop deployments?
- How do I plan for expansion if I require more compute, memory, storage or networking?