November 13, 2002

Conference Proceedings

I've attended several conferences this fall. Each seemed to have a different idea about how to publish their conference proceedings.

Agile/XP Universe published a proceedings which is also available from Springer. It contains the papers from the conference, plus one-page descriptions of the workshops, panels and tutorials. The volume looks professional. They handed this out to attendees when they arrived. I was pleased to get a chance to skim through the papers of interest to me that first night so i could figure out who i wanted to talk to. Slides for many of the keynote presentations were made available (to anyone) on their website after the conference. Many of the keynotes had clearly been updated up to the last minute.

Quality Week provided a hard-copy of the presentation slides and a CD that contained both the slides and the papers, for presentations that had them. A short-coming of this was that it made reading the papers cumbrous during the conference, when i'm often most motivated to read them. Anyone who attended a tutorial got a copy of all the tutorial slides. They also sell copies of the proceedings directly.

The STAR and Test Automation conferences hosted by SQE provide a hard-copy of all the presentation slides. Papers, however, can only be accessed from their Stickyminds website, after the conference. And you have to search the website for each paper: you can't just obtain the proceedings as a single download. This is particularly frustrating to a consultant friend who only has a slow dialup connection from his home office: it makes them unavailable to him. The website includes both the presentation slides and papers. These are available to conference attendees, as well as to their premium subscribers.

The Pacific Northwest Software Quality Conference provides a hard-copy of their proceedings that includes the papers and the slides for the keynotes. They don't collect or distribute the slides for the paper presentations. They also generously make available their entire proceedings for free to anyone from their website.

Conclusions

Paper's need to be made available for reading. It doesn't make much sense to me to have a conference paper that can't be read until after the conference. I like having printed copies handed out at registration. Better would be make copies available shortly before. As a review board member of one of the conferences, i had access to electronic copies of the papers ahead of time and had read through all of them before the conference. I got a lot more out of it because of it; it's too bad the other attendees didn't have the benefit of this.

Some conferences require papers for presentations, others don't. Conferences that are based on papers have presenters who are more aware of other developments in the field. I hate it when someone presents an idea that's been in the conference circuit for years but which they think they invented. This doesn't happen when presentations are based on peer-reviewed papers. This probably accounts for why PNSQC now provides the best testing conference in North America. With most of its attendees from Oregon, Washington and British Columbia, it presents itself as a regional conference. It's not. It rivals the "international conferences."

The hard copies are nice, but they are often often large. I love having proceedings in electronic format, whether they were distributed on CD or in downloadable format. I store copies of proceedings on my laptop so that i have them for reference. I often refer to them at other conferences and workshops -- when online access is rarely available.

Posted by bret at 12:23 PM

November 01, 2002

Rebuilding the Software Testing Hotlist

I believe that this blogging software can be used to host the Software Testing Hotlist.

Each entry would correspond to a link + commentary.

The blog categories would work for the categories of the entries ("Test Automation", "Test Strategy"...). And categories have descriptions which can be used in the indices.

The main page would be organized categorically (and then reverse chronologically), just like the existing site. A separate page would list the most recent additions. This is a big plus from the current layout which makes it hard for repeat users to find new stuff (i'd been trying to keep lists of recent changes manually, but the overhead of doing this is one of the things that's been making me disinclined to update it in the past year.)

I'd been wanting to add more book reviews. I could enter these like article references, except that the "extended entry" could include a fuller length book review.

Entries can be turned back to "draft" mode after they've been published. This might be used in cases where we have dead links, but we don't want to remove the entry from the database for good. Since there are perl API's for everything, it's quite possible that we could write perls scripts that would automatically check for stale links and pull them from the list automatically if they'd been stale too long. This has been a pain for me to handle manually.

I've been thinking about how to create a tool for updating the software testing hotlist for a while. An unexpected benefit of using movabletype would be that i could partner up with others on the project.

Posted by bret at 10:36 PM

How I Draft Entries

Like all my writing, i write the first draft of my entries in emacs and then paste them into the movabletype entry form. Emacs automatically inserts line wraps (because i programmed it that way). So i turn off "Convert Line Breaks" so that the line endings are recomputed by the browser. This, however, requires that i manually insert paragraph breaks ("< p>") between paragraphs. If i make a habit of this, i'll write an emacs macro to do this automatically.

"Draft" entries are in the database (MySQL), but don't appear on the pages. You have to "publish" to make this happen.

Posted by bret at 10:10 PM

Creating at Stickyminds Look-Alike

After reading the user manual and surfing other blog sites, i'm pretty sure that movabletype can be configured to support a site for posting columns -- similar to Stickyminds.

Presentation. It can be configured so that the main index page shows the last "entry" only, including the comments inline at the bottom (in either chronological or reverse chronological order).

Comments. Any one can post a comment. They need to provide an email address, but apparently you can spoof someone else without any trouble. So there is a potential here for fraud. The only way to combat this is by deleting comments, and by blocking comments from specified IP addresses -- pretty weak. On the other hand, it allows people to comment more than once. One more thing: comments can embed paragraph breaks!

Review. Authors can post drafts and then have them "published" by an editor. I see this mostly as a way of checking format, style, etc. I don't see that it requires this review process: except that "authors" might not have publish rights, which would require that an "editor" be the one to publish.

Archives. Much easier-to-use URLs are available for permanent links.

Email. The system allows people to subscribe so that they will automatically be notified when a new entry is made.

Intro. You can have a short blurb ("entry body") that introduces the full column ("extended entry"). This allows you to create indices with just the short blurb.

There are of course other interesting things to discuss: like who would publish, how often, and other editorial matters. I'm more inclined to accept more flexibility in publication frequency (although i'd like to give each column at least a week's run) and column length. In fact, these kinds of editorial matters are what i'd really like to get comments on. I think the technical issues are mostly decided.

Posted by bret at 10:03 PM

Test Automation Frameworks, Carl Nagle

A solid description of the design, purpose and options for data driven testing frameworks. This is part of a larger project to place framework code in the public domain.
Posted by bret at 09:02 PM

Origins of the Software Testing Hotlist

The Software Testing Hotlist started as a personal convenience back in 1995. At that time i had three computers that i used to access the web: two at work (one windows, one unix) and a windows box at home. I had better equipment at work, but access to usenet only from home. Netscape had just come out with support for storing personal hotlists. I wanted to share the same hotlist between my three machines. Netscape stored the hotlist in HTML. The simplest way to share it was just to put it on my publicly available website.

At the time, i had recently decided to focus my career on software testing. I'd been testing for several years, but with the thought that it was just a temporary job until something better came along. My hotlist include whatever i found to be of interest, collecting items and adding annotations so that i could follow-up in more detail later. The web was new, and there wasn't much available about software testing. I mainly collected websites and articles, for further research or citation.

I was just starting to write professionally back then. I presented my first paper in the field at Quality Week in 1996, put it on my website and linked it into my hotlist. Then i started publicizing my hotlist to help attract readers for my paper. I let a couple of search engines know about it and since then readers have grown mostly by word of mouth.

Once it was published, i found that people would send me interesting items to add, which made it easier for me to find new things. I also started notice that many sites were linking to it. When Google later came out, this linking put the hotlist at the top of its list of software testing sites.

Posted by bret at 04:30 PM