A Very Active Repo


Whether you are a new or veteran Pandora owner, you are all familiar with the repository (also known as the repo) available on the openpandora.org site. Recently, milkshake, who is in charge of the repo management, released some figures related to the repo usage since 2011. The results are interesting.

Let’s cut through the blah-blah and go straight to the facts. The repo usage is increasing over time, even though it is lower for particular months. You get three graphs for the price of one. The first one shows the number of repo downloads over time, in months. The second one the overall volume downloaded from the repo over time, in months. Note that that second one is not super reliable since it considers the final size of all software in the repo for the calculation of the volume exchanged: it does not keep historical filesizes. The third one is a cumulative graph showing the total number of downloads over time.

Repo downloads (number) per month since repo release
Repo volume downloaded (in gB) per month since repo release
Cumulative repo downloads per month since repo release

Since the second chart in red is probably biased, I will make a couple of comments on the first and third chart in terms of downloads.

  • First, it’s not like we see a constant progression of the downloads. There is clearly a plateau effect between April 2012 to November 12, where the actual number remains very much stable. 
  • Then, from November onwards, there is a sudden increase, reaching a peak in January, remaining stable in February, and an even larger peak in March, followed by a strong month of April, higher than both January and February.
  • From the third graph, you see that the pace of downloads is accelerating from early 2013 onwards. Actually, more than 30% of the total repo downloads since end 2011 are from the first months of 2013 alone! Pretty impressive when you think about it.
  • Not sure what May will be like, but it might be lower than April if we trust the first 15 days of data.

So, what’s happening here? And, more specifically, what are the drivers for more activity ?

  • More Pandora users? Well, technically, yes, there are more and more of them, and it’s likely than every one of them impacts the repo downloads big time, since they may be downloading all their apps at once versus a regular user who only downloads a few apps every month. But unless I am mistaken, we do not believe there is a spike in terms of sales, at least nothing that should warrant should a sudden increase. 
  • More Popular apps? Probably the strongest driver here. DraStic by itself has been responsible for several thousands of downloads, and may well be at the origin of a few spikes. Many other high profile apps were released during the DragonBox compo that was running earlier this year, so there seems to be a correlation here.
  • More updates? Not only new apps make for new downloads. PNDManager makes is a breeze to update existing PNDs, and these are counted as downloads as well. So, if a few popular apps are updated at the same time, you would very likely get a spike in downloads. Like what, oh my God, you know, PCSXReARMed was updated in January, and this means thousands of people re-downloaded it, and January is a very strong month. PCSXReARMed being the number 1 application on Pandora, it’s very likely that single update is responsible for a strong showing in January. Firefox, updated in January and March. Second most downloaded application. Two strong months. Coincidence? I don’t think so. Of course, it does not account for everything, but they probably add up to the rest quite significantly.
  • New apps? Not really… many apps are poorly advertised by their authors or packagers, and you can see it the low amount of downloads they get. Having 50 poorly advertised apps is not driving downloads very much, no sir.
  • More press? March was the strongest month ever. Is there any relationship with… the launch of PandoraLive in March? !!!! BIAS ALERT !!!! Let’s not pretend that PandoraLive’s importance is huge on the Pandora scene, but several users have been reporting that it made it easy to follow the repo news and further increased the awareness of some applications. It probably counts for something, since there are several thousands of unique visitors per month. If even only half of them download an app because it’s mentioned on PandoraLive, this could account for quite a difference in the repo downloads. We’ll need to do a survey on this one day.

So there you have it. Facts, and some suppositions. It sounds like a good detective story. Most likely there is no single culprit driver for this activity increase. My intuition tells me that in the order of importance… 1) Updates of popular apps, 2) popular apps, 3) more exposure, 4) new apps and 5) new users make a difference.

Why are popular apps behind updates, you ask? Well, even something as amazing as DraStic only gets about 300-400 active users over several updates (leading to 1700 downloads over several iterations). It is probably safe to assume that there are more active users of PCSXReARMed and Firefox, especially since DraStic has yet to progress in terms of emulation speed.

This being said, let’s not say this is smoke without fire. There IS stronger activity on the repo, and as I have shown it a couple of months ago, the pace of apps release is indeed accelerating. It’s all good.

One of these days I will write a piece about basic marketing techniques on how to advertise your new shiny application like a chief. Because, seriously, if you are a developer and spend so much time making a PND, why not spend a couple more hours making people interested in it?

Till next time, and happy downloading on the repo! Comments are welcome, as always.

Leave a Reply

3 Comment threads
10 Thread replies
Most reacted comment
Hottest comment thread
4 Comment authors
nuhrinNeelixekianjosehs33 Recent comment authors
newest oldest most voted
Notify of

A couple of interesting things to note: It’s more obvious in _wb_’s graph on the boards but there is an interesting anomaly around Nov2012, where the bandwidth spikes, even while the number of downloads dips. This spike would appear to coincide with the bulk of the development of the extremely large Code::Blocks PND which saw several revisions during the course of that period. Since the PND in question has grown significantly in size since then it seems likely that the bandwidth figure for November is greatly exaggerated. You mention the plateau beginning April 2012. What’s more interesting to me me… Read more »


” Is there any relationship with… the launch of PandoraLive in March? !!!!”

Yup I believe this is one solid reason! Keep up the great work 🙂


“Because, seriously, if you are a developer and spend so much time making a PND, why not spend a couple more hours making people interested in it?” Good question. It occurs to me that some developers spend an awful lot of time at their jobs on things which are ultimately driven primarily by marketing departments, and that environment can be ambivelent or hostile to the idea that writing software is a craft. Spending so much time making something just for the sake of making it can be a great joy, and spending even a little time trying to “sell” it… Read more »


“Spending so much time making something just for the sake of making it can be a great joy, and spending even a little time trying to “sell” it may be contrary to the spirit and intent of just playing at craftsmanship.” So perhaps a shift in perspective is needed. It’s not really about “selling” your work, it’s about taking pride in it and showing it off. Consider a museum, every piece has an information panel telling what it is and why it’s significant. Without that information it can sometimes be difficult to understand why a piece is on display to… Read more »


I was not advocating for, or arguing against, any particular point of view being better than another, but rather just giving my thoughts on the question, “if you are a developer and spend so much time making a PND, why not spend a couple more hours making people interested in it?” I’m a bit surprised by your reactions to my comments. If you feel that a developer’s work is shitty or prideless because they choose to approach a work for its own sake rather than as a service to as large an audience as possible, well I’m sorry to hear… Read more »