« November 2004 | Main | January 2005 »

December 22, 2004

Tufts hires Amy Rich, Unix guru and Sysadmin Author

A few weeks back our consultant, Amy Rich, notified us that after 7 years of being a consultant with OceanWave she was getting out of consulting to look for full-time employment. It was a bummer for us. She's built all but one of our systems and we'd relied on her expertise for all things sysadmin.

I was talking to her today about finishing some stuff up and she indicated she'd found a new job and was starting tomorrow. Out of curiosity I asked where she was going and she says . . . Tufts :}.

Amy is starting tomorrow as the new senior sysadmin for Tufts' university systems group (USG). I can't say how good it is to have her coming on board. She's been helping I've been working with her ever since I came to Tufts (in 2000), and have been reading her Q/A section in Sysadmin Magazine about as long. Amy's helped us out in many-a-pinch with anything from getting a terminal on an older Unix box to digging into firewall preformance issues.

Welcome aboard Amy, looking forward to (continued) working with you.

Posted by mike at 10:32 AM

December 20, 2004

Building a Simple RPM is . . . Simple

Over the last month or so I've been helping folks at some other schools update or install new versions of the TUSK code. Every installation is Linux; Reh Hat 7.x, Fedora Core 2, Red Hat Enterprise 3. I'm not crazy about RPMs, but they have a purpose and I'm just glad that when they are appropriate they are simple to build.

Part of our install requires making a static build of Apache, compiling in mod_ssl and mod_perl, and tacking Embperl onto the build. Getting all the steps and options just right is a trick, using a script is the only sane way to go. While it's possible to have the script make install, my preference is to build a package. This helps if there is a need to upgrade or an undetected problem with the build, you can confidently uninstall and then reinstall the new package.

I've built close to a zillion packages on Solaris, using the Sun pkg programs, and had installed an RPM or two, but had never built one.

The process is simple, easier than all the manual work required with Sun's packaging.

1 - create an rpm home dir (~/rpm)
2 - inside the rpm home create 5 directories; BUILD, RPMS, SOURCES, SPECS and SRPMS
3 - create ~/.rpmmacros with two lines

%_topdir /home/mkruck01/rpm
%_tmppath /tmp
4 - Download the tarfile you want to install (apache in our case) and put it in SOURCES.
5 - Emacs SPECS/apache.spec (Emacs knows a spec file and creates a nice template)
6 - Add info to template to look like:

Summary: apache web server
Name: apache
Version: 2.0.54
Release: 1
License: Apache
Group: application
Source0: %{name}-%{version}.tar.gz
BuildRoot: %{_tmppath}/%{name}-%{version}-%{release}-buildroot
%description
Apache is a great webserver

%prep
%setup -q

%build
./configure --prefix=/usr/local
make
make test

%install
rm -rf $RPM_BUILD_ROOT
make root=$RPM_BUILD_ROOT install

%clean
rm -rf $RPM_BUILD_ROOT

%files
%defattr(-,root,root,-)
%doc
/usr/*

7 - Build the software and RPM with: rpmbuild -bb ~/rpm/SPECS/apache.spec

The RPM will be in RPMS/i386 (or whatever your processor identifier happens to be).

The cool thing is that for many different software build processes the build and install steps are the same, so you can copy the spec file and change the first few items, drop your tarfile into the SOURCES dir and have an RPM faster than it would take to untar and compile manually.

Posted by mike at 5:55 PM

Installing Fedora Core is more work than Gentoo

A few weeks back I made the decision to replace Gentoo with Fedora Core 2 on a machine used for our wiki. I justified it because I wanted to get the box built quickly and I believed that Fedora Core 2 would be a quick install. I was wrong.

Yes, after downloading the CDs you can use the graphical installer and be up and running in less than an hour. Unfortunately that's not where "building a box" ends for me.

First, I noticed that apache was several versions old, then that FC2 was using openssl.very.old. A new version of openssl wasn't available in up2date (according to a co-worker it's because a newer version isn't trusted yet - but the old version with several known exploits is better?). I decided, since the reason I'm rebuilding this machine is because it was hacked, that I need to be on a newer version of openssl.

The rpm dependencies show a ton of stuff that relies on openssl, I started to work through them and after 30 minutes got sick of it and just went ahead and build a new RPM and installed. Of course, as soon as I did that SSH failed. I went over to the physical machine and thought I'd get a new SSH built. Without openssl, wget dies looking for the libraries so I figure the best solution is to start over.

I rebuilt the box, choosing to install nothing except the basic packages. Everything I wanted to use (apache, perl, php, MySQL, openssl) were all significantly old enough that I'd rather have the barebones OS and build everything myself. I download openssl, openssh and wget. Next build and install a new RPM for openssl, and rebuild ssh and wget. After the good part of a day, I'm finally on my way to moving past the OS and onto the webserver, database etc.

Isn't this one of the main reasons I went to Gentoo, to get away from depending on RPM availability, the out-of date packages and having to deal with the constraints of RPM dependencies. How did I forget all that annoyance.

The truth is that I wouldn't have had to rebuild the box in the first place if I had been better about using Gentoo's emerge with some regularity, or at a minimum do a pretend and email the results.

Since we're up and running, I'm going to stick with FC2 for the time being, but next chance I get it will be back to Gentoo for that box.

Posted by mike at 5:00 PM

December 12, 2004

First few Sections of Book Written and Submitted

This morning I sent off a few sections of my first chapter of the book to the publisher and co-author for a look. Feeling a little vulnerable and curious about how it will measure up.

In a conversation with the co-author on Saturday we agreed to send each other what we'd written so far, ccing our coach at the publisher. Although I was completely unsure how my writing would stack up, it is a great relief to know that someone else has a chance to tell me why it stinks (assuming that it does). The idea was to let each other take an early look and make suggestions.

The last time I did any significant graded writing was a paper on Leaders in Technology for my master's degree, back in 2000. Yea, I write this weblog but I don't spend a great deal of time on making sure the writing is clean and conveying the right message.

Posted by mike at 11:22 PM

December 10, 2004

Open Source Financial System for Higher Ed

Along the lines of Linux-like movements, I just recently heard about the Kuali Project, which is an attempt to replace PeopleSoft, Oracle or whatever a school uses for it's financial system.

While Sakai ruffled a few feathers and is getting some attention from commercial vendors, I would imagine that Kuali will recieve a lot more resistance. The financial stakes are much higher for commercial financial systems as the revenue is significantly more than a LMS license.

Posted by mike at 11:26 AM

Is Sakai the Linux or Apache for Higher Ed?

One of the themes of this Sakai Conference, which was proposed at the opening keynote by Brad Wheeler, is that the development of Sakai has the feeling and potential of the Linux movement.

I've been thinking about that all conference, there is a lot of energy and excitement about using Sakai, but the real need is to contribute to testing, discussions, building tools etc. The Linux and Apache projects wouldn't have gotten very far if everyone thought it was a great idea but didn't commit to contribute. In talking to folks here there are some who are anxious to get back and start moving, others express that they are sceptical and nervous and don't want to commit.

As for us at Tufts, we came in thinking this would be something to watch, but we're starting to think that we might reallocate some of our development efforts and contribute to Sakai. The idea is that we would net a lot if we worked on moving parts of our system which aren't features of Sakai. Once our tools work in Sakai we could replace our core with Sakai, and open up umpteen possibilities for collaboration with other schools. In some cases we could get a full-featured tool with little work, like the impressive gradebook tool Berkely has developed. A gradebook has been on our list for a long time, but it's such a monumental task we haven't been willing to commit.

I guess you can't really predict a movement like Linux or Apache, and maybe you jinx a movement by saying "we're going to be a major movement." We'll see in another 6 months, when the Sakai schools gather again, how much has been accomplished.

Posted by mike at 10:48 AM

December 9, 2004

Jump Around with House of Pain on Shared iTunes

I'm starting to realize that the only bigger time-sink than a used CD store for me is a nearby iTunes Shared Music folder. I've yet to come across someone who doesn't have a collection worth working through. Tonight, it's 837 songs from ben in the W New Orleans on the 23rd floor.

Yea everyone's got a few random "blast from the past" tunes that are fun to play for a minute, but so far in every anonymous sharing experience I've had there have been at least one or two CDs that I genuinely have thought about buying and can really enjoy.

Jump Around really isn't my style of music . . . really. And no, I won't tell you how far it has climbed on my Top 25 Most Played songs.

Posted by mike at 11:35 PM

Implementing Sakai in a Community College Consortium

Listening to Vivian Sinou from Foothill College talking about their process of implementing Sakai in a consortium of colleges.

Small group, Vivian asks for introductions.

eTudes (French for study) developed in house in 1995, started with a faculty member offering a course via email and getting in trouble. 48 partners schools using eTudes, mainly in California but also in a few other states. As Foothill College has wanted to migrate to Sakai they've had to think hard about all those partners.

Foothill applied for a grant from HP who came back and said they should join the Sakai effort, 2.5 million to join Sakai.

Steps after Getting the Grant
- gap analysis
- hiring/resource allocation
- learn new technologies
- gather requirements
- establish project plan
- communicate with core
- monitor progress closely
- migration strategy
- user support
- launch spring quarter pilots

Foothill is creating a tool for authoring lessons, includes a ton of functionality. Can be integrated with samigo, includes IMS content format import/export, show/hide courses.

The Foothill College authoring tool will be released with Sakai 2.0.

Posted by mike at 2:40 PM

Sakai Tools Team Update - Style Guide Overview

Am listening to a presentation from the tools team at Sakai Winter 2004 Conference on style guide. A new style guide (1.3) is being released at this conference. The 1.2.2 styleguide is up now. Sakai takes consistency and accessibility very seriously.

Style Guide
Work in progress which describes standards for Sakai development. Looked at all core schools CMS tools and design/usability guidelines and principles. Styleguide includes design and behavior requirements for tools.

Structure of Style Guide
- what's new
- preface
- written materials
- views (descriptions and sample)
- ui components

Why use the guide?
- consistend user experience across the appliction
- tools are used across institutions [and different skins]
- consistend mental models
- relieve cognitive load

The Usability Professionals' Association
has a Designing the User Experience poster which is a nice representation of the process of designing for the user experience.

How to use the guide
- get famiar with the document
- use it early
- UI changes made in style guide
- follow page layout and widgets
- language consistency
- informs basic interaction

Building a new wireframe
- start with an existing view (in Visio or OmniGraffle, files provided by Sakai)
- take the parts you need and drag them into a new diagram
- modify it to fit your design

Use labels for interaction behavior, CSS stylesheet coding guidance etc to elimintate confusion during conversations about the design.

All new 1.5 design work and 2.0 design work are using a common CSS stylesheet.

Posted by mike at 11:53 AM

What's New in Sakai 1.5?

Rob Lowden, who leads the Sakai tools team, is presenting a session at Sakai Winter 2004 conference about what new features are coming in Sakai 1.5.

In many places on the site Sakai is using HTMLarea.

Personal Profile
- view users in the system
- edit user profile (may be pre-populated from directory)
- add photo
- create HTML body bio

Syllabus
- create a site with multiple syllabus items - with WYSIWYG HTML editor
Context Sensitive Help
- comes with static help for existing Tools
- for each tool in the system a help system is available so if people are looking for help on a certain page the help that comes up is targeted to that tool
- submit a help form/trouble ticket

Calendar Widget
- anywhere in system a date is needed there is a little calendar widget

Site Management Tool
- search for sites by date, school and department
- administrators can "adopt" a role when using the system, to help see system as another user for testing, debugging
- reuse, duplicate, copy sites
- homesite layout - create views that appear on all site pages

User Management
- search
- edit username, roles and status
- identify manyally added participants (not created in bulk)

SAMigo
- assessment authoring
- assessment publishing
- assessment taking
- assessment grading
- question pool management
- template management

Gradebook
- integrated with SAMigo
- coming after 1.5 release, but before 2.0

Coming in 2.0
- advanced resources tool
- sessions-bases syllabus tool
- worksite overhaul (redo to match latest UI spec)
- help authoring

Posted by mike at 10:32 AM

December 8, 2004

iTunes thinks R.E.M. Lyrics are Explicit?

I did a search in iTunes tonight to see if they've got R.E.M.'s latest CD and the results startled me. If you haven't listened to R.E.M. before, they aren't explicit as the iTunes seems to think. I thought for a second that the songs were from a live CD or something but even then couldn't imagine Stipe doing anything that would make one, let alone all of the songs require this warning.

Didn't end up buying the new album, I've got physical copies of all the others, want to be able to look at the cover design and liner notes.

Posted by mike at 11:23 PM

If You want to Develop Sakai Tools

If you want to develop tools for Sakai, the technologies you should be experience with:

- JavaBeans
- Servlets
- Interface design
- OKI OSIDs and Sakai APIs
- Maven
- JavaServerFaces
- Hibernate

Hmmm. I know what all these are, not sure that qualifies as experience.

Posted by mike at 5:28 PM

Sakai Educational Partners Conference gets Started

Got into New Orleans late last night and am sitting in the first session of the Sakai Educational Partners conference.

If you haven't heard of Sakai, it's a joint effort by a few core schools to create a Collaboration and Learning Environemt. What does that mean? Most people are using something like Blackboard, WebCT or homegrown software to provide a service to faculty, students and staff for creating and viewing content. Sakai is attempting to create open source software which does CMS, LMS, CLE (or whatever you like to call it) better than it's ever been done, considering requirements.

The Sakai Educational Partners Program allows non-core schools to be a part of the conversation, and shape the direction of Sakai.

One question I like to ask of fellow attendees is what their Sakai roadmap is. Most often people are involved to be sure their needs are voiced, and to keep an eye on the software to determine at what point they might migrate from their existing system to Sakai.

From my experience, Sakai is very thourough. Everything is done in slow-motion with lots of conversation about use cases, requirements, standards etc. Probably the way every project should be done, but sometimes taking the time to fully flush out all the issues isn't realistic.

The conference roster contains ~200 folks from ~70 institutions, mostly universities in the US and Canada.

Posted by mike at 10:50 AM

December 4, 2004

External DVD Burner for iBook

I've been considering getting an external DVD burner for my wife's iBook, which I bought without the Superdrive (I was short-sighted, she's constantly coming to me with iMovies or iPhoto slideshows wanting a DVD burned).

A little looking around and I find a likely candidate, the Formac Devideon. As I'm digging around to find out more information on the software I realize that Formac just slapped a slick case on the Pioneer DVR-108 and bundled it up with their DVD-authoring software and are selling it for $199.

I do some searching for the DVR-108 and see I can get it for $74 at Newegg, or I can go to CoolerExpress.com and get the burner in a Bytecc IDI->Firewire encloser for $120. Add in Roxio Toast Lite 6 for an extra $22. Roxio Toast Lite is Toast Titanium without the extra frills like network burn, vinyl record conversion.

That price makes it more compelling, still not 100% convinced. Part of me is excited to see if DVDs authored by something other than iDVD and the Superdrive are compatible with more DVD players, but another part of me wonders about the quality of the Roxio DVD-authoring software.

Posted by mike at 9:27 PM

December 3, 2004

Deciding to Author a Book

Two months ago I got an email from a publisher inquiring whether I'd ever considered authoring a book on a particular subject (to be revealed in a future post).

For the first day I was in a bit of a daze, not knowing where to start asking questions. I decided since I didn't even know what questions to ask it would be good to take some time just to put together the questions. I wrote back to the publisher and said "I'll get back to you in a week, after processing."

Heidi (my wife) was the first to contribute to the question list and offer suggestions. One of my primary concerns was how pouring time into writing would affect our family. Still a little concerned about that, but Heidi assured me we could find a way to clear time for me to write.

My quest to generate a good list of questions then went to Pete, who thought one of the most important things I could do was talk to other authors and have them share both thoughts on relationship and arrangements with the publisher as well as observations in the authoring experience.

Over the week I got information (varying anywhere from email to IM to phone conversations) from my father, Phil Windley, Jim Fazio and Jeremy Zawodny, each offering excellent snips of advice.

The list of questions:
1) Am I interested in writing?
2) What are the reasons I'd want to write?
3) Do I have time?
4) Do I have the knowledge and experience (or drive to find the knowledge and get the experience) to write on the subject?
5) Is there room for another book on the subject matter?
6) Is the press respectable?
7) How does the press treat it's authors?
8) What does the contract look like?
9) What's the advance, royalties?
10) Should I use an agent?
11) Are there co-authors? Who will find, choose, approve?
12) Could some of the writing be outsourced?
13) How does my work on the book relate to Tufts University (my employer)
14) What are the clauses in the contract for a rejected manuscript or if the press decided not to publish.
15) Am I really prepared for the work (which I won't be able to imagine until it's upon me)?
16) What is the schedule, time frame?
17) Who decides the contents of the book?
18) What are the steps in going from initial idea to table of contents to writing chapters?

With this set of questions, some of the personal ones answered, I went back to the publisher and agreed to go forward and have a dialogue about the questions I couldn't answer on my own.

More to come . . .

Posted by mike at 12:03 AM

December 2, 2004

Winter Time - Hot Chocolate Time

I've been drinking a lot of hot chocolate lately. I think because when I get up in the morning it's cold in the kitchen and the thought of a warm cup to start the morning helps me through getting breakfast ready for the kids.

I got this mug from Pete back in 2001 for Christmas, he made it at one of those do-it-yourself table-ware places. It's big enough to justify two servings of hot choco mix, but not quite big enough for two cups of water which makes for a thicker drink. A little spurt of whip cream, the kind from the atmosphere-damaging can, adds an extra punch of sugar.

I like this kind of hot chocolate.

Posted by mike at 8:55 PM

MySQL 5.0.2 Alpha Out

Wandering around on the MySQL site and I see that 5.0.2 is now available, with binaries for most platforms. There is a warning about it not being a real release, still having some outstanding critical bugs. I've been building from the BitKeeper tree for the past little bit to be up to date with the latest on my experimental server, will have to give the 5.0.2 packages a go.

Posted by mike at 8:16 PM