The QuinCe Alpha is being taken offline in preparation for a new version to be released. This is a complete rewrite for both the internal and user interface, with support for many new features that we found were missing from the first versions. More details will come when it’s ready.
Alpha 1.08 has just been released, which fixes a major bug that frequently prevented QC changes being saved in the system. Thanks to the people who reported it – it’s taken some time to work out what the problem was, but it should behave much better from now on.
Also in this update is a change to the colour scale on the maps, which should make them more useful.
The latest new feature for QuinCe has just been added – you can now view your data on a map in the QC page. A new Map button has been added under each plot, which will show you something like the above. As with the plots, you can choose which variable to look at – just press the “Variable…” button. The colour scale is automatically adjusted to the 5%-95% range of the data (to prevent outliers making the scale meaningless).
The map can be scrolled and zoomed just like the Google Maps you’re probably familiar with. Additionally there are four buttons visible. + and – are standard zoom buttons. The S in the bottom right shows and hides the colour scale for the current data. The Z button will reset the map to show the whole data set in view (the same as what you see when the map is first displayed). Hovering over a point will show its position and value in the bottom left of the map. Clicking a point will locate it in the table just like the plots.
You’ll will notice that as you scroll and zoom around, the points being displayed will change slightly. QuinCe only loads 1,000 points into the map at any one point to help keep the performance up, and as the view changes it loads as many points as it can in the current view. This means that single outlier points may not always be visible in the map as they are in the plots, but if you zoom in close enough they will appear.
The colour scale could use some improvements – I’m trying to find something that’s good for colour-blind people, but that usually involves blue which isn’t so good for the ocean! I’ll continue experimenting with this. Eventually I want to highlight flagged points on the map as they are for the plots. Development on this will continue. There are all sorts of other things that are possible – synchronising the plots and maps so highlighting a value on one also highlights it on the other, for example. Any other fancy ideas are welcome, and I’ll consider them as time allows.
As always, any and all feedback is appreciated.
Alpha 1.06 of QuinCe is running as of this morning. This contains a set of minor bug fixes that caused issues in the database – there are no new features in the user interface, but hopefully it will be a little more stable.
QuinCe Alpha 1.05 has just gone live.
Multiple Row Selection
The major change here is the ability to select multiple consecutive rows using shift-click. It should be fairly intuitive, but here’s some quick instructions anyway.
- You can select a row by clicking on it.
- You can deselect a selected row by clicking on it.
- After you click on a row (either to select or deselect it), shift-clicking on another row will select/deselect all the rows up to and including the row that was shift-clicked.
- There have been some minor changes to the comment entry dialog to make it work better with the new selection mechanism.
- Your session should no longer time out if you stay on the QC page for a long time.
- Processing of longitudes in the western hemisphere has been fixed. The bug shouldn’t have affected your experience so far.
I would like some feedback on the way the data is loaded on the QC page. At the moment, the base page is loaded, and then the data for the two plots and the table are downloaded. Only a portion of the data in the table is downloaded at any one time, to reduce the delay when the page first loads. As you scroll around the table new data is loaded on demand.
I am concerned that loading the table data on demand is taking too long. For me it’s more or less instant, but that is probably because I’m on the same network as the QuinCe server. I’d like to know how it performs for people at further distances. Could you try this out and send a rough indication of (a) how long it takes to load the initial QC page, (b) how long it takes to load a new chunk of the table data, and (c) whether you think this is too slow (especially for point b).
The alternative approach is to load all the data for the QC page up front. Doing that will mean that the initial page will take much longer to load, but after that it should run more quickly. I may develop a version of the page that does that to see how it works differently based on feedback.
This is a small release to fix a couple of obvious bugs:
- Dates are displayed properly on the QC screen
- Row highlighting works more reliably (I’m not confident enough to say fully reliably…)
- Some minor improvements to the WOCE flag dialog
There were several issues with user session handling, including:
- Being told your session was expired when it wasn’t
- Getting a blank screen after logging in
- Getting a blank screen after logging out
These issued have been fixed.
There have also been fixes for sending confirmation emails, which frequently failed.
Fixed a bug in the QC page that prevented the automatic QC flag being overridden with a Good flag.
Alpha 1.01 is a small update to increase the reliability of the login page. Functionally, there are no changes to the site.
There is still an issue with the logout link that just goes to a blank page. The logout works fine and ends your session properly, but it doesn’t go back to the login page as it should.
The first version of QuinCe is up and running. This is a proof-of-concept release, to show the process of uploading data files and performing quality control. The site is not fully functional, and will use a restricted set of data simply to give an overview of the system to a few volunteer users. The purpose of the release is to ensure that there are not any fundamental concepts that have been overlooked before development continues.