Today, my main task was to make a new release of the WebPA files. This should be a simple job, but our timing is really off. Just recently Sourceforge have moved their systems to a new data center. Which seems to have had a effect on some of the services that they provide, such as the wiki and the FTP upload for project releases.
This has prompted conversations about reporting on our view of sourceforge and the issues and limitations that we have found with the system. I have in the past written about the 'sourceforge' effect and perceived views and values of the system, which in a way covers the issues that we are facing at the moment. But with all things in the open source arena, they are providing a free to use service so some give and take needs to be afforded.
For us, this is a glitch that we will just have to cope with for now and hope that they will have the systems over at Sourceforge up and running again really soon.
Showing posts with label sourceforge. Show all posts
Showing posts with label sourceforge. Show all posts
Thursday, 19 June 2008
Monday, 3 March 2008
Other JISC projects - open source, sustainability ... the lot
Following my post last week, related to whether it is feasible for WebPA to become sustainable and the impact of the funding, a post popped up on Rowins blog on essentially the same topic.
Rowin's post is about three JISC funded assessment tools. The three tools though have left becoming open source till their funding has ceased. So what is the future for these projects, well I don't know. I fear that, although there are seven developers and a sourceforge area, there is no wider community to support the future of these projects. Further funding would possibly extend their lives, but as pointed out time and time again, with no community to support the projects, is there a future in the short term worth paying out for.
I would love to know from all of the JISC funded projects, funded at the same time as WebPA;
I know that WebPA is the exception to all of those, but at times it has been a fight to get every ones buy in to the open source ethos and way of doing things. Some of the barriers and limitations on to the project have been discussed in this blog. I hope to continue raising these sometimes awkward points, at least till the funding ends for us too, or in the perfect world we reach sustainability.
Rowin's post is about three JISC funded assessment tools. The three tools though have left becoming open source till their funding has ceased. So what is the future for these projects, well I don't know. I fear that, although there are seven developers and a sourceforge area, there is no wider community to support the future of these projects. Further funding would possibly extend their lives, but as pointed out time and time again, with no community to support the projects, is there a future in the short term worth paying out for.
I would love to know from all of the JISC funded projects, funded at the same time as WebPA;
- how many have released their code under open source,
- how many are beginning to build up a community to support their tools,
- who has waited till the end of the funding to release their tools,
- has any one really shouted loud enough to gather interest in the tools, to build the community's,
- is there any tool which is likely to reach sustainability.
I know that WebPA is the exception to all of those, but at times it has been a fight to get every ones buy in to the open source ethos and way of doing things. Some of the barriers and limitations on to the project have been discussed in this blog. I hope to continue raising these sometimes awkward points, at least till the funding ends for us too, or in the perfect world we reach sustainability.
Friday, 16 November 2007
Measuring our sucess to date
After the release of the WebPA OS download we are happy to report that a total of twenty four (24) downloads have been made since the beginning of October 2007. We are unable to gather statistics on where these downloads are from, which is a shame, but it is encouraging that others are interested in the WebPA tool .
The reason why we are unable to gather more information than the download numbers is due in part to our use of sourceforge. Sourceforges privacy policy prohibits us from being able to access more detailed information. It a way it is sufficient to know that there have been twenty four (24) downloads of the WebPA tool. However, it would be nice as a JISC funded project to be able to identify the UK downloads. I don't know that the selection of another open source community tool would have made much difference as to the stats available, but there may have been a bigger impact on the download level achieved (possibly negatively).
Further to the downloads, we are happy to be able to also report the success of the WebPA Demonstrator (passwords). Since its launch on the 29 October 2007, the demonstrator has been used by people in the following countries; Australia, Sweden, New Zealand , and of course the UK. Within the UK we have had visitors from; Loughborough, Hull, Derby, Belfast, Oxford, London, Bristol and Edinburgh to name a few (from 17).
I have been able to collect more information about this by using the Google analytics tool. From the collection of the statistics, it is possible to evaluate how long users have stayed on the site, and in some cases presume why they leave so early. But there is no way to see how the use of the demonstrator links to a download of the WebPA tool from sourceforge.
The last interesting point is that WebPA has already reached a global audience. Although this is beyond the remit of the original funding, at the same time it can be seen as quite gratifying. Especially for those at Loughborough who have been involved with WebPA from its original inception.
Hopefully in the future I will be able to return with more information on how the WebPA project is progressing in terms of downloads and demonstrator use.
The reason why we are unable to gather more information than the download numbers is due in part to our use of sourceforge. Sourceforges privacy policy prohibits us from being able to access more detailed information. It a way it is sufficient to know that there have been twenty four (24) downloads of the WebPA tool. However, it would be nice as a JISC funded project to be able to identify the UK downloads. I don't know that the selection of another open source community tool would have made much difference as to the stats available, but there may have been a bigger impact on the download level achieved (possibly negatively).
Further to the downloads, we are happy to be able to also report the success of the WebPA Demonstrator (passwords). Since its launch on the 29 October 2007, the demonstrator has been used by people in the following countries; Australia, Sweden, New Zealand , and of course the UK. Within the UK we have had visitors from; Loughborough, Hull, Derby, Belfast, Oxford, London, Bristol and Edinburgh to name a few (from 17).
I have been able to collect more information about this by using the Google analytics tool. From the collection of the statistics, it is possible to evaluate how long users have stayed on the site, and in some cases presume why they leave so early. But there is no way to see how the use of the demonstrator links to a download of the WebPA tool from sourceforge.
The last interesting point is that WebPA has already reached a global audience. Although this is beyond the remit of the original funding, at the same time it can be seen as quite gratifying. Especially for those at Loughborough who have been involved with WebPA from its original inception.
Hopefully in the future I will be able to return with more information on how the WebPA project is progressing in terms of downloads and demonstrator use.
Monday, 15 October 2007
Measuring the success of WebPA
Last week I attended a JISC workshop on evaluating JISC funded projects. Although a lot of the evaluation seemed to be steered towards the student experience, it did raise a number of points related to evaluating the project. These are especially valid for those projects where a large proportion is related to the development of software.
In the past I have sat in no end of meetings where evaluation has been discussed, and the software is quite often overlooked, or in many cases the evaluation of software is not understood. I have seen no end of project submissions where the evaluation for the software produced is out lined in terms of "the code will be peer reviewed" but as an opensource project that will happen over time and will be carried out by the community that develops for the project rather that a nominated project or organisation. Anyway getting back to the software the evaluation should be in relation to how widely the system is now used and the attitudes of those using the system.
One phrase that came out of the workshop which has stuck with me is "Metrics for Managers". I think this is a good place to start to highlight the types of information that can show the success of WebPA (after all it will be a success). For those who are interested in what WebPA can do for the institution we need to look at the student retention rates, rather that the rather softer student experience. By evaluation the retention rates we may be able to show that with the use of WebPA for peer assessment, the retention rates increase. However, this may be far in advance of what can be done in the time remaining for the delivery of this project.
There are other measures that we can use for the evaluation of WebPA, link the ones that I have highlighted in the past with out even thinking about evaluation. We are already able to collect data related to the download rate of WebPA, we just need to be able to marry this back in to the number of Institutions across the UK and then by sector that have adopted WebPA. This is more difficult that it may seem to begin with. For a start yes we can find out the number of downloads for the software we are receiving (12 to date), but due to the privacy policy for Sourceforge I can not find out any information about who is downloading. This then leads on to the second element, how will we know who is adopting WebPA either on an institutional level or individually in their teaching. The only what of measuring this will be the metrics that we can gather from other areas of the sourceforge system and form contact that we have with institutions and academics. But I do think that these metrics will be in some cases difficult to obtain and also miss out some vital information, however, they will be useful anyway.
Even though evaluation is normally left till the end of the project I think the JISC timing for this event was appropriate. I think that it will trigger project teams to start thinking about what they can evaluate and how, in turn this will enable the teams to start collecting data sooner, rather than scraping around for information and trying to make it fit towards the conclusion of the project funding.
In the past I have sat in no end of meetings where evaluation has been discussed, and the software is quite often overlooked, or in many cases the evaluation of software is not understood. I have seen no end of project submissions where the evaluation for the software produced is out lined in terms of "the code will be peer reviewed" but as an opensource project that will happen over time and will be carried out by the community that develops for the project rather that a nominated project or organisation. Anyway getting back to the software the evaluation should be in relation to how widely the system is now used and the attitudes of those using the system.
One phrase that came out of the workshop which has stuck with me is "Metrics for Managers". I think this is a good place to start to highlight the types of information that can show the success of WebPA (after all it will be a success). For those who are interested in what WebPA can do for the institution we need to look at the student retention rates, rather that the rather softer student experience. By evaluation the retention rates we may be able to show that with the use of WebPA for peer assessment, the retention rates increase. However, this may be far in advance of what can be done in the time remaining for the delivery of this project.
There are other measures that we can use for the evaluation of WebPA, link the ones that I have highlighted in the past with out even thinking about evaluation. We are already able to collect data related to the download rate of WebPA, we just need to be able to marry this back in to the number of Institutions across the UK and then by sector that have adopted WebPA. This is more difficult that it may seem to begin with. For a start yes we can find out the number of downloads for the software we are receiving (12 to date), but due to the privacy policy for Sourceforge I can not find out any information about who is downloading. This then leads on to the second element, how will we know who is adopting WebPA either on an institutional level or individually in their teaching. The only what of measuring this will be the metrics that we can gather from other areas of the sourceforge system and form contact that we have with institutions and academics. But I do think that these metrics will be in some cases difficult to obtain and also miss out some vital information, however, they will be useful anyway.
Even though evaluation is normally left till the end of the project I think the JISC timing for this event was appropriate. I think that it will trigger project teams to start thinking about what they can evaluate and how, in turn this will enable the teams to start collecting data sooner, rather than scraping around for information and trying to make it fit towards the conclusion of the project funding.
Monday, 8 October 2007
Getting there slowly
I am in the process at the moment of supporting a number of institutions with their own installations of WebPA.
The first institution is well on their way to getting the installation complete. However, I am not to clear on what version/revision of WebPA they are using, so supporting any queries that come in is interesting, as I can not tell if I have previously fixed the problem.
The others are also on their way to getting an installation, which is really exciting. But again, the issues that come through the door are always directed to me, rather than through the WebPAsourceforge area . This leaves me to add the information to the trackers myself. At the moment this is okay, as there are not to many requests. I am encouraging the users to use the sourceforge system, so every time I email back with the solution I endeavor to include tracking information, as well in, the hope that at some point they will use the system.
The other major success that has occurred over the weekend is the number of downloads of WebPA we have had. I released a version of WebPA as a download on Friday 5 October, since then we have had five downloads. To me this is good going, but only time will tell if it is a success.
The first institution is well on their way to getting the installation complete. However, I am not to clear on what version/revision of WebPA they are using, so supporting any queries that come in is interesting, as I can not tell if I have previously fixed the problem.
The others are also on their way to getting an installation, which is really exciting. But again, the issues that come through the door are always directed to me, rather than through the WebPAsourceforge area . This leaves me to add the information to the trackers myself. At the moment this is okay, as there are not to many requests. I am encouraging the users to use the sourceforge system, so every time I email back with the solution I endeavor to include tracking information, as well in, the hope that at some point they will use the system.
The other major success that has occurred over the weekend is the number of downloads of WebPA we have had. I released a version of WebPA as a download on Friday 5 October, since then we have had five downloads. To me this is good going, but only time will tell if it is a success.
Monday, 6 August 2007
Sourceforge an adventure into the unknown
Although I have used Sourceforge for a number of years to locate applications and also find fixes to the said applications, I tended to be a list squatter, only interacting anonymously. But now with this project I am having to learn all the nuances of the systems and how to administer them.
Even though Sourceforge may not be the best repository to put our project into it is one of the most well know, as I have already mentioned in a previous blog post. So it is better the devil you know etc... My first delve into Sourceforge was a number of months back now when I registered my self and then the project. This simple straight forward and painless. Sourceforge stated that it may be hours to days before the project was registered, in our case it was accepted and live within hours of the submission. Feeling really chuffed I didn't really touch anything else for a bit. Yes I know that if the project is to be successful the more activity needs to take place, but hey I'm one developer on a funded project, with enough work to do without sitting on Sourceforge.
Anyway, my next delve was to set up a list for feature requests and set up the few other lists for people to use when the project moved forward again. This was still a painless process... But just last week I decided that I really should be using Sourceforge to store all the code, so I bit the bullet and began. In my past roles I had used CVS, and at home we set up SVN so how hard can it be! Well this was just the beginning of the pain and some self inflicted!
After diligently sitting and reading through all of the instructions on Sourceforge, and then down loading and reading the instructions for tortoisesvn. I tried my first upload of the information, which didn't work! I went through all of the information again trying to find out what I had done wrong. Eventually I worked out my error, as I didn't read one element of the document, mainly due to the statement at the top of the section that said new users of the subversion system did not need to read it. However, the answer to my problem was within this section...the URL structure!
It was a little after my initial jubilation of getting the code on to Sourceforge that a colleague pointed out that I had not edited the config file. At this point I had uploaded a couple of URLs and passwords that I did not want public. This is when the real pain started...
I knew from reading the documentation that deleting anything was going to be difficult. My next set of pain started closely after when my brain let me down and my historical knowledge of *nix machines reared its ugly head. Here I was sat at work scratching my head, I had managed to make a connection with putty, and back-up the repository, I just couldn't work out how to get the subversion admin filter to work, or locate exactly where I was in the file structure. At this point it seemed best to go home and get my hands on all the better tools I have there.
Once at home I soon followed the instructions, and then requested Sourceforge run the data import for me. As with setting the project up on Souceforge this took no time at all. But I did get a error...after rereading everything again I was trying to run the import on files which where not in the correct place. Once this was figured out then the files where moved and the import request process run again. This time everything was successful, for the import at least, but I had just uploaded all the incorrect data again. I went through the information again, worked out how to get the svn version I wanted, and went through everything again, this time with the successful out come that I required.
Once I got passed the fact the documentation is mainly written for people who use systems like this, day in day, out and expect anyone using the system will have the same knowledge and experience, it is actually straight forward. Administering the rest of the site has been fairly easy once you start thinking the way they want you to. But I do get the odd moment, like this morning when I lost a request that had come in. I have since found it again, but only through starting again.
If you are interested at looking at the Sourceforge area for the project, you will find it at: http://sourceforge.net/projects/webpa/
Even though Sourceforge may not be the best repository to put our project into it is one of the most well know, as I have already mentioned in a previous blog post. So it is better the devil you know etc... My first delve into Sourceforge was a number of months back now when I registered my self and then the project. This simple straight forward and painless. Sourceforge stated that it may be hours to days before the project was registered, in our case it was accepted and live within hours of the submission. Feeling really chuffed I didn't really touch anything else for a bit. Yes I know that if the project is to be successful the more activity needs to take place, but hey I'm one developer on a funded project, with enough work to do without sitting on Sourceforge.
Anyway, my next delve was to set up a list for feature requests and set up the few other lists for people to use when the project moved forward again. This was still a painless process... But just last week I decided that I really should be using Sourceforge to store all the code, so I bit the bullet and began. In my past roles I had used CVS, and at home we set up SVN so how hard can it be! Well this was just the beginning of the pain and some self inflicted!
After diligently sitting and reading through all of the instructions on Sourceforge, and then down loading and reading the instructions for tortoisesvn. I tried my first upload of the information, which didn't work! I went through all of the information again trying to find out what I had done wrong. Eventually I worked out my error, as I didn't read one element of the document, mainly due to the statement at the top of the section that said new users of the subversion system did not need to read it. However, the answer to my problem was within this section...the URL structure!
It was a little after my initial jubilation of getting the code on to Sourceforge that a colleague pointed out that I had not edited the config file. At this point I had uploaded a couple of URLs and passwords that I did not want public. This is when the real pain started...
I knew from reading the documentation that deleting anything was going to be difficult. My next set of pain started closely after when my brain let me down and my historical knowledge of *nix machines reared its ugly head. Here I was sat at work scratching my head, I had managed to make a connection with putty, and back-up the repository, I just couldn't work out how to get the subversion admin filter to work, or locate exactly where I was in the file structure. At this point it seemed best to go home and get my hands on all the better tools I have there.
Once at home I soon followed the instructions, and then requested Sourceforge run the data import for me. As with setting the project up on Souceforge this took no time at all. But I did get a error...after rereading everything again I was trying to run the import on files which where not in the correct place. Once this was figured out then the files where moved and the import request process run again. This time everything was successful, for the import at least, but I had just uploaded all the incorrect data again. I went through the information again, worked out how to get the svn version I wanted, and went through everything again, this time with the successful out come that I required.
Once I got passed the fact the documentation is mainly written for people who use systems like this, day in day, out and expect anyone using the system will have the same knowledge and experience, it is actually straight forward. Administering the rest of the site has been fairly easy once you start thinking the way they want you to. But I do get the odd moment, like this morning when I lost a request that had come in. I have since found it again, but only through starting again.
If you are interested at looking at the Sourceforge area for the project, you will find it at: http://sourceforge.net/projects/webpa/
Tuesday, 31 July 2007
What has happened to July
OK we are on the last day of the month and usually by this point I will have posted as many as four posts, but as you can see this month is some what lacking. Well there is a good reason for this, as many of you know me will know that I have been really ill over the past few weeks. I seem to be on the mend again and have made it back into work.
However, nothing has been static on the project! In my absence Paul Newman has stepped in and more or less saved the day. As a result we have been able to send off a copy of WebPA to the University of Derby. The University of Derby are our first pilot institution. As I find out how they are doing with the installation of the system I will of course post information to here.
Getting this far has not been an easy road though. Some of the major barrier have not been to do with the systems code but rather with PHP itself. One problem in PHP is related to the inclusion of files. Within WebPA we use 'require_once' which is built on include. However, their is a problem including files with the relative path. As WebPA needs to be cross platform this has been a serious issue. PHP has been corrected for the latest release, but we as a project can not guarantee that everyone will be using this. At the moment we seem to be able to make thing work. At present I have WebPA running on the MacBook Pro, but not within the local user sites area. Once I have this fixed, I will try the code on the LAMP server and then on a machine I have set up at home running WAMP. Once I am certain that the system is working on all the architecture configurations that I can easily lay my hands on then I will make a public release available via the projects sourceforge page.
As for sourceforge, I'm still not convinced that this is the best location to serve the projects downloads from (see an earlier post). I am watching the Simal project as they are using the Google code tool and I can always ask them how it is going. But as of yet I am still to find an alternative (apart from hosting a suite of tools here at Loughborough).
One other thing that has happened in my absence is that the project has visited another potential pilot institution. However, as expected we are faced with some institutional barrier. These barriers include; academic uptake of the tool, availability of server space and skills to implement the system locally. Due to this we may need to look at approaching other institutions with in England, where there is less of a barrier to adoption. For now though I will have to return to doing my day to day work and get the system out, as we need to 'release early, release often!'
However, nothing has been static on the project! In my absence Paul Newman has stepped in and more or less saved the day. As a result we have been able to send off a copy of WebPA to the University of Derby. The University of Derby are our first pilot institution. As I find out how they are doing with the installation of the system I will of course post information to here.
Getting this far has not been an easy road though. Some of the major barrier have not been to do with the systems code but rather with PHP itself. One problem in PHP is related to the inclusion of files. Within WebPA we use 'require_once' which is built on include. However, their is a problem including files with the relative path. As WebPA needs to be cross platform this has been a serious issue. PHP has been corrected for the latest release, but we as a project can not guarantee that everyone will be using this. At the moment we seem to be able to make thing work. At present I have WebPA running on the MacBook Pro, but not within the local user sites area. Once I have this fixed, I will try the code on the LAMP server and then on a machine I have set up at home running WAMP. Once I am certain that the system is working on all the architecture configurations that I can easily lay my hands on then I will make a public release available via the projects sourceforge page.
As for sourceforge, I'm still not convinced that this is the best location to serve the projects downloads from (see an earlier post). I am watching the Simal project as they are using the Google code tool and I can always ask them how it is going. But as of yet I am still to find an alternative (apart from hosting a suite of tools here at Loughborough).
One other thing that has happened in my absence is that the project has visited another potential pilot institution. However, as expected we are faced with some institutional barrier. These barriers include; academic uptake of the tool, availability of server space and skills to implement the system locally. Due to this we may need to look at approaching other institutions with in England, where there is less of a barrier to adoption. For now though I will have to return to doing my day to day work and get the system out, as we need to 'release early, release often!'
Subscribe to:
Posts (Atom)
webpa