So I tried at home on my favourite Windows XP VM and it installed fine. Fine is of course subjective because it required three restarts to actually install itself (one for .Net 4.0, one for Windows Installer 4.5 and the final for Visual Studio Web Developer Express) – I’m not quite sure why these couldn’t have been bundled together into one restart (which is reasonable) or even what VS Web Dev Express was doing to require a restart. But I’ll live. I fired it up and did a Joomla! install and everything worked nicely. I tried the VS Web Dev integration and found it loaded everything up except that Web Dev doesn’t do PHP – back to the drawing board. As an aside why Microsoft doesn’t support PHP in VS Web Dev directly is a wonder – PHP as a language appears to be more popular than either of Microsoft’s according to langpop.com’s view of the world. To be fair, langpop.com is a mildly biased view towards open source areas (Google Code, Slashdot, IRC, Ohloh, Freshmeat) which would explain the dominance of languages such as PHP or Python. In any case, it worked and it had a funky editor with highlighting and not a lot else as well as a database editor that handled MySQL – and all in one package.

So I try to install it at work. First hurdle: I need Windows XP SP3. Ok, bummer. File a job, wait for ICT Service Delivery (yes that is their real department name) to get back and install SP3. They call me back in a week and say “when do you want this? is now a good time?” and I say sure. A few hours later and 64+ updates my computer has SP3 on it. Let’s try Web PI 3 beta again and see how we go. Much better, we’re off installing WebMatrix and all sorts of shiny. I ticked the Joomla! box as well and that is also installing. Everyone is happy. It then goes to try and configure IIS on my XP box. It pops up a component install view window and eventually gets to the point of asking me for an XP install CD. Except this is a corporate box and I don’t have one. I don’t have an i386 directory handy to just copy files from. I have local admin so installing stuff isn’t an issue but I don’t have those files. Bummer. I go hunting for them and I can’t readily find them locally and I don’t even know where to find them on the network. Sheepishly I hit cancel and WebPI keeps installing without configuring IIS 5 on my local machine. I thought I’d selected to use the WebMatrix server so I’m not entirely sure what is going on anyway. .Net 4.0 installs uneventfully (no restart this time) and Windows Installer 4.5 requires a restart. This time I’m not installing Web Dev so I don’t need to restart for that, yay! The rest went relatively smoothly which is good however at the end it complained IIS5.1 failed thus items dependent upon it (e.g. FastCGI for IIS5.1, PHP5.2.13, WinCache 1.1 for PHP 5.2) failed to install. MySQL Windows Essential 5.1 didn’t work because a service entry couldn’t be created, possibly because a MySQL service entry already exists. Joomla! didn’t work either but since both IIS 5.1 and MySQL failed to work I’ve got a selection of options.

Curiously I had an XAMPP install of MySQL set up and WebPI didn’t detect this properly opting to install its own version, perhaps a check on port 3306 to see if something speaks MySQL would be more appropriate to check. Improvements for next time.

So I fire up WebMatrix Beta and use its installer to install Joomla!. This one works better. It presents me with the same dialogue and insists I can’t have a blank password (my insecure XAMPP MySQL install insists otherwise but I secure it and both my security office and WebMatrix are happy). Joomla! appears to install perfectly fine now. Go figure. It also appears to have used the XAMPP MySQL install which is nice.

So some quirks in no particular order:

• Requests has an images filter which is nice but if you enable it, go from requests into something else and back again the filter button is depressed but the list isn’t filtered. Clicking on it filters the requests but messes up the state.
• Requests has an images filter which is a good start but no JS filter or CSS filter.
• The built in DB tool connects to MySQL fine however if you try to create a table with a column as both an “identity” (Microsoft SQL Server version of “auto_increment”) and a “primary key” it complains about multiple keys. Setting the field to just be identity appears to resolve the issue
• Perhaps most annoyingly the font choice is harsh, at least for me. Compared to Monaco on my Mac, it just looks disgusting. Unfortunately there doesn’t appear to be a way to change it either. After coding for half the morning in Smultron whilst my desktop was updating, WebMatrix was just garish and jutting in comparison.
• Whilst there is no autocomplete for the PHP as far as I could see, it would be nice to include a reasonable formatter. At least something that fixed indentation by default. Mind you, Smultron doesn’t have this but I do miss it at times.

I also at one point managed to get WebMatrix to terminate however you need to remember that WebMatrix is still in beta as a product so it is understandable that there might be some issues that will hopefully get fixed. All in all it looks like a good tool. As a cut down editor it seems to fit the bill quite nicely as well as being one of the easier ways to get up and running on Windows with a web development environment and PHP.

I went home, spent most of the afternoon crafting my assignment together. And then submitted it. To do this I need a TurnItIn originality report. Easy enough. I do all of that, wait for the report to come up, download the report (0% copied!) and upload it and the assignment to EASE. I have the sneaking suspicion that I had forgotten something and re-read the specification. I had, it was an introduction and a conclusion. I was close to writing one but didn’t because I forgot to put it in the original outline (I had everything else mind you). So I write up an introduction paragraph and a conclusory paragraph, regenerate the PDF (using LaTeX, NeoOffice on Mac is just a hog, is slow and non-responsive. TeXShop on Mac is fast and lets me write what I need with minimal fuss. Random style is hard but such is life) and resubmit to TurnItIn. Since I’ve already submitted they’re telling me I will need to wait two days. This really annoys me that I need this to submit, my university is paying anyway plus if we used the Moodle API we wouldn’t have this problem. In some respects it is a form of extortion because they want a material support to release it in some respects (signing up again with a new account would obviate the problem). In any case I submit anyway with the old report. We’ll see how we go.

Just recently I had issues with Outlook Web Access doing stupid things. I think my mailbox was over quota because when I emptied the trash via Apple Mail, everything started working. The weird thing is that when I clicked “Check Names”, it nuked the entire email for some reason. Curious behaviour. Anyway, email sent with mild frustration. Another day passes.

So I had the on going issues with the wiki. Some appear to be legitimate, others appear to be an inability to copy and paste accurately. This is rather unfortunate but such is life. They are only first year students, so hopefully they’ll learn at some point. There were a few that weren’t in this bucket and I managed to get them resolved. We also had one student who changed their email address to a non-USQ student email address and is complaining that they didn’t get any email notification. I have a distinct feeling that the email was dropped as the student had a Yahoo! account and they have a habit of blacklisting us and our emails don’t go through. This is in part why we have the UMail accounts so that we can avoid this problem. We can’t blacklist ourselves.

Other issues involved some hosts needing to be added ezproxy’s list. It appears that EBSCO were redirecting this particular resource back onto ezproxy even though it was free. The ezproxy upgrade we did appears to have had this strange side effect that I will need to spend some time investigating if we can get rid of it so that for URL’s it doesn’t recognise it will unset itself.

The next issue was a link in the journal list that wasn’t being rewritten through ezproxy. That was an easy matter of updating the link where it was to run it back through ezproxy. I’m not entirely sure how it got changed but it was changed none the less and fixed now.

I had the regular Thursday tasks that the person off sick would have done. This involved kicking off the MARC export of our Library Management System, VTLS Virtua, and then starting the 11 hour reindex process into Solr and VuFind, the tool we use to index the catalogue. This bit doesn’t take much time however I spent a bit of time working on improving some of the scripts. I altered the way it handled logging and printed some user friendly messages to the screen. I’ve made small improvements along the way so hopefully this will continue.

Another curious thing was EndNote again complaining about a particular page implicating them more than VTLS. The issue is that since the upgrade EndNote has been broken. The original text read like this:

APOLOGIES – PLEASE NOTE: Currently there is no connection file for the USQ Library catalogue. We are awaiting a compliance upgrade from EndNote to match our Library Management System. We apologise for the inconvenience this causes. As an interim measure you might choose to search the UQ or other Australian Libraries for the title once you’ve found it in a normal websearch of our catalogue. We ask you to understand that this is NOT a permanent situation. Notification of the availability of the USQ connection file will be made ASAP

EndNote were upset about the implication that their system could be at fault here and that they had to do something. Curiously EndNote is the only system impacted and malfunctioning. Every other system that utilises the Z39.50 interface works perfectly well including but not limited to our Library Catalogue, the desktop management tool and web search tool provided by VTLS, OCLC’s VDX, Civica Aurora, Biblioscape 8, Yaz and Mercury Z39.50. Both Yaz and Mercury are freely available tools and Mercury was even tested off campus by a colleague at UQ. There are perhaps other systems that interface on Z39.50 that I have missed but it looks very much like the only system that isn’t working is EndNote. This has now been changed to the following:

APOLOGIES – PLEASE NOTE: Currently there is no connection file for the USQ Library catalogue. We are awaiting a compliance upgrade of our Library Management System to fix this. We apologise for the inconvenience this causes. As an interim measure you might choose to search the UQ or other Australian Libraries for the title once you’ve found it in a normal websearch of our catalogue. We ask you to understand that this is NOT a permanent situation and is not a problem created by the EndNote software. Notification of the availability of the USQ connection file will be made ASAP.

Fortunately VTLS have accepted that they made a change that broke EndNote and they’ve said they’ll fix it in their latest revision. It is unfortunate that EndNote seem disinterested in resolving the issue when VTLS took ownership and I must admit I never got a reasonable response from anyone about what was happening. As far as I could see from the logs, Virtua issues a challenge to EndNote which never gets responded to and EndNote just times out on its own. So whilst VTLS has agreed to do a fix to resolve the issue and restore compatibility however it will be in their next release which will be in a few months and due to the time it takes us to complete formal testing around the system it will likely be another few months after that until the fix reaches our production system. Unfortunate really. What I really don’t get is why EndNote is the only product that isn’t working but such is life. VTLS are happy to fix the problem and it’ll get resolved eventually. To be honest I don’t care who fixes it, I’m just disappointed that it took us a long period of time to get anyone to fix it.

Another day slowly turns.

Happy Easter!

I visited with ICT’s senior security officer, Luke, to work on configuring the load balancers to make library.usq.edu.au work better than it is presently. It still isn’t perfect but we’re a while ahead. Now when you go to library.usq.edu.au it doesn’t redirect around to libcat.usq.edu.au (well not immediately anyway) but serves up the page which is nice. Eventually it’ll behave like this normally but until we get there this is what we’ve got. The next step is easy enough now we’ve got this set up. We’ve also now got the library blog (http://library.usq.edu.au/blog) using /blog instead of /wordpress, shifted it to a new server (now hidden by the load balancer as well) and also performed an upgrade to ensure it was on the latest version of WordPress. We ran into an issue with firewalls on one of the servers making life difficult but one of our sick admins, Jayish, saved the day and fixed up the issue with the server. Jayish from home also granted more RAM to our library wiki box that we’re using for engineering problem solving and updated the internal and external DNS for the library load balancers. We got everything sorted just after 8am or so which was good.

Next trick of the day was to update the library wiki to remove the dependence on the LDAP server. Foswiki, the software that runs the wiki, took the approach of trying to sync the entire LDAP directory seemingly on every request. I think I’ve configured it correctly but I’m not entirely sure. To be honest whilst this is the first time I’ve seen this style of doing things (e.g. trying to suck in the entire LDAP directory), in my experience with JAuthTools I’ve had it suggested to me a few different times that the LDAP integration should work that way. It doesn’t particularly scale well and in our situation we’ve got our directory but we can’t limit it down to the particular courses because of the indexing that is in our LDAP directory. At some point we might get the indexing to work but even then its going to put an unreasonable load on the LDAP servers to handle the searches as well. David, one of the other fellows in Library Systems, did a lot of the hard work to create the new htpasswd file with all of the entries for each of the users we have in the system and handled the email notification. Fortunately everything appears to have gone well which is good – and no significant complaint emails or phone calls. Yay!

The middle of the day progressed towards meetings and emails. I had a meeting with my boss and his boss for a short while to confirm some stuff, spent a bit of time with the social club for their morning tea and had a chat with our resident Mac system administrator from ICT about our Mac. Some how this bit of non-productiveness sucks more time than it should, which is unfortunate. I ended up having lunch, a tad late but I took longer than usual. After 1pm was a meeting with ADFI about the media repository. More work that I’m going to need to do to get everything connected up and running which should be fun in some respects. I had a quick wander around Y block whilst I was there and had a chat with David, Matt and Jonathon. I had a chat with them about the changes I made to DiReCt the other day and Jonathon pointed out that I didn’t need to file a technical issue, I could have a chat with one of the USQ functional support staff, Warwick, to get it updated. So I wandered down stairs and had a chat with Warwick and we got it sorted. This will hopefully mean that students from the UAE (predominantly what distinguishes the students with issues) and others on weird ISPs or limited work accounts (I don’t think even USQ would let you in if it wasn’t on campus already) won’t be seeing the weird timeout or connection error messages that they have been seeing. One more thing resolved for the day. Whilst I was on the second floor I went for a wander to haras Jason, one of the sound guys but he was busy – or at least pretending to be anyway, it is rather easy when you’ve got signs up.

That pushed me into the mid afternoon and a few more meetings to round my day out because you can never have enough meetings. I finished off by demonstrating JIRA to my boss in an attempt to get him to use it more. Might even pan out. I rebuilt a dashboard for him with custom filters and all sorts of mildly relevant information. We will see what happens.

I finalised the day by cleaning up on some documentation, actually clearing off some of the paper on my desk (or at least sorting it into nice piles) and updating my infrastructure todo white board to note the changes I made this morning (things on the list still include resguide environment refresh, moving the podcast system and some extra dev environments). Might write some more documentation tomorrow, for something new.

I finished up and went to uni to have a meeting with my supervisor. We went over the paper that we’re working on to get through a few things. I had a good two hour lunch catching up with people. It appears that ICT is going to be outsourcing with the people currently being sued by the University of South Australia. A curious choice, to go with someone who has a track record of apparent failure but they were probably cheap enough. I followed this by a short visit in my office to catch up on things before I went to visit with Luke, the ICT Security Manager.

We managed to get the library test site and blog working properly behind the load balancer which unfortunately took around 45 minutes to get right. Whilst I was visiting, he also fixed up coursereading pointing to test not prod which is good. This means I need to double check a few more things and we’ll be ready to

Ok, so I deleted everything in my temp folder, restarted and saw what happened. I saw the cnc4launcherupdate.exe file being created which was funky but then I received the same error message. So I deleted it, relaunched and got the same error message. Eventually after doing this repeatedly the launcher eventually loaded. Clearly you don’t a) check that the file is properly downloaded and b) prompt to redownload in this situation (given that you don’t check it to begin with it would be rather hard to do).

Once the launcher loaded I was stuck on the r1core_data_cnc4.game.rtp file not downloading properly and crashing. I am still stuck at this point.

So eventually got an update to this:

Hello there,

Thank you for contacting EA Games Support,

I sincerely apologize for inconvenience caused to you. I have checked the DXDiag of your computer the drivers are out dated please install the latest driver. In order to know your drivers open the dxdiag file and go to the “Display Devices” and check the drivers under ” Driver Date/Size” You will find how old your driver is. To correct this you have to update your computer driver, I have included the link for your convenience please go through and make necessary changes.

Should you have any further questions, please take a moment to review our Knowledge Base site at http://support.ea.com.

Sincerely,

EA Rep Austin
Electronic Arts

I’m annoyed that they’re apologising and not actually reading my responses. Drivers might be a problem but I initially a) couldn’t launch the launcher and b) the launcher can’t patch properly so it doesn’t matter what version of my display drivers I’m on, I’m not even getting there.

Seriously?

I mean, seriously? Your launcher has a flaw in it that it doesn’t validate that the download has completed successfully (even a simple MD5 sum or file size check should fix this) and then tries to run the app which fails because the file isn’t downloaded correctly.

Problem number two is what appears to be an issue that the launcher when you finally get the damn thing updated hands on downloading the r1core_data_cnc4.game.rtp file. This doesn’t appear to be an uncommon issue looking through the forum.

Don’t worry, I don’t need an answer to this. I gave up. I wanted to play the game and I got a crack. The game runs nicely and I’m progressing through the missions fine. Thanks for not reading my messages. Repeatedly.

The funny thing was that whilst I uninstalled and reinstalled the game I got a CNC4 server emulator downloaded, copied it into place, made the required changes to my hosts file and I started playing the game. It is working perfectly fine. It depresses me that I got what I couldn’t have paid for – I could have illegally downloaded the game and I would have actually been playing it faster than the legal game. As it stands the legal version isn’t still working anyway so in some respects I’m still behind by paying money. So I responded to their message with the text above and I also ticked the ” I no longer need an answer to this question.” I don’t require a response any more because it is working, just not their intended way.

Of course the sad thing is that the game is working fine and I’m enjoying it. It is a different concept that I’m trying to get my head around but it is pretty enjoyable. I haven’t had much time to go through it and I’ve only done the first few GDI missions but I’ll get there eventually. It looks like there are a few glitches which could be patched but since patching the game legally doesn’t work I guess I’ll never know.

I had some fun with paperwork and a rather unfriendly colleague trying to be annoying. Managed to get all of the paperwork filed and shifted around the place which is good. It all ended up working out even though it was at the last minute. Nothing a bit of friendliness doesn’t help.

I was having a look at getting the JIRA jelly to notify non-users from JIRA when their issue changes. We don’t necessarily want people to be able to do anything but just be notified of changes. Creating users for everyone would seem a bit painful instead of just adding a few emails to the particular issue so that they’re notified when things happen.

To round out the afternoon I spent some time redoing the MARC loads for a new type of file and hopefully that will get properly tested tomorrow and all sorted.

Late in the afternoon I had an email from another part of the University. They wanted to attend some of the sessions to get an idea about the architecture about DiReCt since we’ve got two consultants coming up to try and work out how the system works. What I don’t get about this is that they hired the person who knows the most about the product at the University and gave them an extra two levels on top of what they were being paid. I don’t quite understand why need to get an idea when they already employ an expert already. But such is life.

The work slowly turns!

The next meeting was a pseudo meeting for myself where I filled in a few more HEAT requests for the day. I had three, including one requesting to install HEAT at the suggestion of Luke Drury. I’m told it is only used in ICT related business. Such is life, I guess Service Desk really enjoy rerouting all of my HEAT jobs.

Following this was a meeting with one of the Faculty Librarians about Access All Databases (AAD) and looking at options. AAD is an application that has been picked up from elsewhere and repurposed by the Library and it shows. It is in dire need of a rewrite and it scares me how badly the entire item is structured. We’ve got some changes and we’ll have a demo at the relevant meeting later in the week. Lucky us! More work likely but no time to clean it up properly.

Next on the list was the ePrints meeting, the last for the quarter. We did alright this quarter and most things were completed acceptably. There are some outstanding items and a few things left but all in all it wasn’t too bad. Was a relatively quick meeting which also makes me happy as well and I got to lunch at a reasonable hour instead of being horribly late.

After lunch I ended up doing some quick support and investigation as to why something wasn’t displaying properly in our library catalogue. Looks like a data error but you can never be cetain about these things. We’ve changed the data but it will take a while for everything to flow through due to the design of the system.

From here the next meeting was to work with ePrints with an ICT staff member to implement a feature that I had previously built into ePrints. Whilst I was working on the Author ID project within ePrints one of the needs was to reference a data set that was stored and managed outside of ePrints, namely the department tree. To do this I created a new field type called the “ExternalItemRef” which extended from the built in “ItemRef” field type. ItemRef links ePrint datasets together and allows you to do references which is cool, ExternalItemRef does the same thing but instead of using an ePrints defined data set, it works with any particular table that is available to the system within the same database as ePrints. You define the name of the table, the key for the table (a unique integer value) and then a text value for it that is then displayed. This gets processed by the field and neatly integrated into ePrints. Further to this I have a widget that permits the user to have a popup dialogue displayed to enable searching of item results. The popup dialogue was itself new functionality that I added to the system to enable searching on the various tables easily. Because it is a popup it can present information in my mind a lot easier than a simple AJAX drop down list. The thing that took the longest was in fact loading the data from an Excel spreadsheet into the MySQL database but after a bit of work it was easy enough to do and we’ve got instructions for next time. The trick appeared to be not using Excel but OpenOffice.org – free software to the rescue. Also regular expressions are amazing at data transformation. I need to do some more practice at loading data into MySQL and work out more effecient ways of doing it but I feel I’ve got it figured out enough for this situation.

The afternoon was spent reviewing the various systems that are available internally, looking at their issues and working my way through my horrific inbox. So many little things that need doing, cleaning up and fixing. One day at a time!

Whilst I was typing this up I ‘timed out’ a few times. This wasn’t in a screen far in, it was the first screen after selecting the session time. At this point the only persistent information that could possibly time out is the movie, date and time. All values I feel should easily be persisting at this point. There should be no issues with those values timing out. The trick is that the timer is for the entire booking session. You have 8 minutes to pick how many tickets you need, what extras you want and your payment method. I also managed to trigger a runtime error as well within the application, a well written application indeed.

The booking process, when it isn’t crashing, asks for a member number as well. It says that is has already filled it in for me but then works out that it really isn’t the right number. So I proceed without it which is unfortunate. But that is the story here.

The process for all of this acquires an email address, credit card number of other identifying item. And that is the key. That is the value in this transaction that they should be encouraging.

The process doesn’t necessarily make it easier for their front counter staff beyond having everything in the system and being able to go through slightly quicker with the order pre-done. It is about building a database of information. Not only can you register people with CineBuzz to get them to provide details and link it in but even if they don’t do that signing in online with credit cards will allow you to handle and work out all of this information into your databases directly. You don’t need to recover it from point of sale systems located all the way around the country, you can have it deposited directly into your system and providing you information. In some cases where you get the booking days before you end up making interest off the money you just took. And you haven’t even had to provide the service!

Sitting through a business intelligence course, I know the value of having information available to you. Encouraging online booking can integrate into so much more. All of the different systems, availability of historical and trending data that is available. Even mild predictive values systems based on online sales and ability to prepare systems. The more information the better. It is worth more than the $1 per person that they seem to be charging for the pleasure of booking online.

Beyond this I didn’t do much beyond a few emails, a small amount of fixing of code and helping people out. To round out my day I did some debate adjudicating and had teams that were starting their debating career in the last year of high school debating. They gave it a go whichis the most important part.

So for the last three months after I got back from my US trip I have started a new position within USQ. Instead of being an Analyst Programmer for the Division of ICT Services (or just “ICT”), I’m not a systems co-ordinator for the Division of Academic Information Services (or just “DAIS”). This is a new position for me which involves a whole heap more meetings that I was doing previously and I don’t have the ability to do as much coding any more and that portion of my ob has been unfortunately replaced with direction and meetings that I asn’t previously tasked with. I do still have a portion of coding and I also have an aspect of system administration in a sense more than I had previously. Some of the systems that I worked on as an analyst programmer were Library related so now I’m the owner of these systems a bit more than I was earlier.

So now within the Library I seem to be tasked with the following items:

• ePrints – more management than development that I was previously involved with
• Access All Databases – more development, this is a new system for me
• Finding Information Tutorial – this is a custom CMS that was made that I am hoping I will be able to port to Joomla! instead
• Library Catalogue Search – another new system, the unified search system for the Library which is pretty cool
• Library Blog – various versions of WordPress, almost all of them out of date
• Mobile Services – a new research project in general for the Library which is fun
• Cleaning up all of the inconsistencies in the services
• A few other things that I’m sure I’ll complain about when I get to them

In the short time I’ve been here, we’ve done a library management system upgrade and accordingly I’ve had to update the library catalogue search. Fortunately in this case most of the work was done so it was a matter of putting parts together to make it all work. I’ve also managed to replace four virtual servers off a legacy virtual machine box that was about to die (four RHEL5 VM’s rebuilt entirely and data transfered in a single day, yay!), recovered a system that was hacked and reconfigured it to meet the start of the semester. I’ve also rebuilt the internal library systems development support environment helpfully called “libtrac”. It is now running Subversion, Trac and a new JIRA instance that I’m using to bring everything together.

Throughout the period I’ve done the typical support tasks, diagnosed issues and in some cases been able to resolve systems. I’ve managed to fix some problems for people but some others I haven’t been able to resolve yet but I’m getting there. One step at a time. The new tracker and Subversion items will allow me to do more items and record things without having the chance to lose them. I’ve also started shifting operational systems into Subversion to allow control of them to ensure that we don’t lose something along the way and we have a consistent methodology for ensuring systems get migrated between one system to the next.

In February I managed to go to my first conference for the Library to VALA 2010 in Melbourne. Even then I managed to present a session for Joomla! which hopefully brought new people into the fold. I’ve also met with different people and learnt a whole heap of things which is pretty cool in the grand scheme of things.

I’ve also had a new staff member (I now supervise two people, scary!) start and they’re now working on the mobile services project primarily as a developer. My other staff member I’m hopefully going to work on skilling them up to be a better programmer than they are previously. I’ve got to get in line with the University’s “BUILD” system for goal setting and performance management. I do wonder if the data goes into a system somewhere or if it is just being dropped somewhere.

Coming up for me is a complete rebuild of some of the oldest of the library’s systems. This includes upgrading horribly vulnerable WordPress installs to better versions and keeping them up to date and upgraded in future. It also involves replacing ancient CentOS boxes and replacing them with newer RHEL5 boxes to get everything almost on a similar version to make everyone’s life easier. This means that the library catalogue system will be rebuilt from dev through test and prod. I’m also shifting “Access All Databases” the other way, creating a new production environment and cloning that back through to a test and dev environment as the production system has issues with availability and SAN presentation problems causing the file system to go read only. Lots of other little things, and small tasks but the world is slowly moving.

Here at USQ we have cases where we have departmental staff in faculties or research centres that deposit papers on the behalf of academics. These might be secretararies or research assistants who handle the initial entry of data into the system, attaching the paper and the other tasks that are required. At the end of this process the library wanted to change the depositor from this user and reassign it to the primary author of the paper. So they needed a way to reassign the depositor. They didn’t want this to occur immediately and wanted to have some control over the reassignment. Initially I built it with a user picker for the depositor to select the new user however the Library requested a generic text field instead that the depositor could use to type the details of the person to which they wished the depositor reassigned.

So the first step in all of this drama is to add a few fields. We ended up with three fields: the first was a field for the depositors to enter the name of the user who the item should be reassigned to, the second was an ItemRef field which linked to the user table to have a representation of the user and the last was a checkbox which controlled if the user of the item is reassigned. The first field is visible to any one in the repository who has an account whilst the last two are visible only to repository staff (e.g. editors or administrators).
So to do this, we need to navigate to /opt/eprints3/archives/[archivename]/cfg/cfg.d/eprints_fields.pl and add the following:

Once we’ve done this, you need to run “/opt/eprints3/bin/epadmin update_database_structure [archivename]” to add the new fields to the database.

The next task we need to do is add this to the eprint workflow so that the fields turn up on the screen. This can be achieve by going to “/opt/eprints3/archives/[archivename]/cfg/workflows/eprint/default.xml” and adding it into one of the stages. I picked the “Notes” stage or tab to put it in, eventually it looked like this:

If you don’t have a “notes” stage already, you will need to add a new one to the flow at the top, if you’re using a standard ePrints configuration, it will look similar to this:

Pretty easy! The next step is to add some translations in “/opt/eprints3/archives/[archivename]/cfg/lang/en/phrases/eprint_fields.xml”. Add these lines before the final </epp:phrases> line:

So there is are two phrases for each field, the name and the help text that goes with it. Most of it is pretty self explanatory. The final step in this trip is to alter “/opt/eprints3/archives/usqep3/cfg/cfg.d/eprint_fields_automatic.pl” to set it so that it will change the depositor of the item once the tick box is checked.

# Depositor Reassignment
if( $eprint->get_value( “new_depositor_update” ) eq “TRUE” && $eprint->get_value( “new_depositor_assignment” ) != 0 )
{
$eprint->set_value( “userid”, $eprint->get_value( “new_depositor_assignment” ) );
$eprint->set_value( “new_depositor_assignment”, “” );
$eprint->set_value( “new_depositor_name”, “” );
$eprint->set_value( “new_depositor_update”, “FALSE” );
}
# End Depositor Reassignment
my $type = $eprint->get_value( “type” );
if( $type eq “monograph” || $type eq “thesis” )
{
These nine lines surrounded by Depositor Reassignment is inserted where it is and at this point we’re done. We’ll need to restart the server before we get too far just to ensure that everything takes properly and then we can test it all out.

The workflow at this point is that a user of the system marks down the person they wish to reassign as the depositor. We have an overnight report which runs to email one of our repository editors to notify them of items that are being requested to be transferred. They then view these items and enter the user ID of the user in question into the new depositor field (they can use lookup to validate it) and at this point they need to check the box to actually update the assignment. In ticking the box the the depositor is set to the value of the new depositor field, the reassignment field is blanked, the new depositor field is blanked and the check is cleared.

The most unfriendly aspect of this is finding the user in question. To do so you need to traverse the user search, find the user and then grab their user ID and put it in the field. This is a bit much. We also wanted to have the option to send an ameila but the editors need to be able to customise the email before it is sent and wanted control.

Fortunately the solution to the first problem was to utilise the Itemref custom InputForm Component that I created for the Author ID project. The second was to create a new one that primitively created a “mailto” link with a lot of the details pre-populated. For the latter I might have been able to do it within the workflow language but I felt it was easier to build in Perl.
Before we begin, we’re going to need a few files:

• The “Itemref.pm” file needs to be put in the “/opt/eprints3/perl_lib/EPrints/Plugin/InputForm/Component” directory.
• The “USQ_DepositorEmail.pm” file needs to be put in the “/opt/eprints3/perl_lib/EPrints/Plugin/InputForm” directory.
• The “user” file needs to be put in the “/opt/eprints3/cgi/users/popup/user” directory. The “popup” directory might not exist in your installation and will probably have to be added first.

Once you’ve got these files in place we’re ready to do some more modifications. In the “/opt/eprints3/archives/[archivename]/cfg/workflows/eprint/default.xml” file where we had the following:

We’re going to change that to the following:

This then tells the system to add a new “Find Entry” button which will popup with the user search window. The user search is limited to repository users so anyone who isn’t authorised will receive a login request form. It also tells ePrints to use the new Itemref InputForm Component to handle this field as well, which is what puts the Find Entry button and handles the other logic.
To add the email option, we add a single line after the above mentioned on to load it up:

Pretty simple changes here so far – this then puts some text with a link up if we need it. The last change we need to make is to include this into our language files. There are a few strings, we’ll just put them into the “/opt/eprints3/archives/[archivename]/cfg/lang/en/phrases/eprint_fields.xml” file we used previously:

All pretty simple, again these lines should be inserted before the </epp:phrases> line in the file. Once you’ve done this, restart the server again to pick up all of the changes and you should see a user picker interface and a little bit of text you can click to make a quick email message. You can customise the email message by altering the language strings to change what you want to display.

And generally speaking, “large, deep-pockets organization(s)” are no more stable than a startup, from the grunt’s perspective. At any moment, you’re one spreadsheet away from being laid off to improve the quarterly statements.

It’s taken from a comment on the Slashdot article about the JRuby guys splitting from the Oracle/Sun behemoth. It is an interesting statement and something I feel is a rather accurate representation of the way life really is and a quite straight forward and direct statement of fact.

Yay for Macroblogging!

But most interesting today I heard word from my old Council friends that they’ve finally gotten to ditching Novell. Whilst I was there a project called “Get Rid of Novell” and “Get Rid of Legacy” which seems to have been given a heightened priority since Novell decided to double the licensing cost of their software. Council now has 90 days in which to remove all Novell software. This should be interesting as they have a lot of Novell supplied software: SLES, SLED, the Netware servers, eDirectory and its client, Novell’s IDM product, iPrint and  Zenworks. One of the things that kept Novell software around was the fact that it was half the price of the Microsoft equivalent, however with this it appears that this isn’t going to be the case so one of the reasons for retaining Novell, cost, is gone. And whilst it is a great solution for a lot of problems, the software that is slowly being deployed within the organisation is increasingly dependent on Microsoft’s software: running on Windows, integrating with Active Directory and providing management tools that integrate into Microsoft’s management console tools. Should Novell had let them be at the original price for another year, they’d have the money instead of doubling the price and well ending up with nothing. But life moves on.

In other news last night I wrote a quick WebDAV interface for Joomla! – its a slow thing and it also works like a bit of a composite as well with supplying an interface into the filesystem. It is using SabreDAV to handle the DAV interface into PHP which is cool but I’m wondering if it is slowing things down.

So with some admin tasks and meetings about different things including Camtasia Relay and ugly administrative tasks. The afternoon was spent working on UniHIRTS. I added a whole heap of code to basically add a workflow check to ensure that incidents or hazards can’t be closed when there are open actions. Something I thought would be really simple turned out to be a pain. But I couldn’t work out why it didn’t appear to work. I ended up putting it at the lowest layer possible and working backwards to find out why it wasn’t being caught earlier. It turns out that there were two functions next to each other that achieved effectively the same thing, except one didn’t run through any of the data validation checks. So once I found that I managed to update the code to use the same code path as the other storage code path that validated the data. Unfortunately in doing this it took me all afternoon digging through the layers of code to find where it actually was after I’d managed to get it partly and badly fixed. Another day in the office.

I spent the afternoon planning for JAOO with Sally and Jon which was fun. Sitting down deciding which sessions we were going to do and how that works. It looks like a really good event to go to and I’m excited to be attending. It also expresses the confidence that the organisation has in me, something I don’t want to disappoint on delivering.

I spent some time trying to work out why some times a before insert trigger on one of the tables for UniHIRTS didn’t work properly. The closest I got was an Oracle “mutating tables trigger error” which sounds sort of close to what the trigger is doing but doesn’t feel right. Every so often the trigger fails to update the composite primary key for the entries which means we need to manually do it, the mutating tables sounds right but we don’t have a record of the error in the log files to work out why it isn’t working. It does seem strange that majority of the time the database updates perfectly but in some situations it doesn’t, especially when the item in particular should be executed in the database.