WordPress 3 puts Pressure on Drupal 7
Drupal users everywhere is eagerly awaiting the release of Drupal 7. The release of WordPress 3.0 two days ago is definitively going to put a little extra pressure on that.
The new WordPress release has naturally generated a lot of buzz all over the Internet and many have already upgraded to it. Even if the upcoming release of Drupal 7 is often mentioned in articles about it, it will still give WordPress the momentum for now. Not until Drupal 7 is ready it will take it back.
Even though the Custom Post Types, flexible menus and custom taxonomies in WordPress core still has a long way to go before it comes close to the much more mature features available for Drupal, it will be enough for many website projects. In most cases those projects would have used Drupal instead.
The longer it takes before it is ready, the more projects will opt for WordPress 3. Simple as that.
Projected Release Date Pushed Forward
There is no release date for Drupal 7, the rule is still that the number of critical issues has to be zero. Right now that number is 65 and for most of us it is going down painfully slow.
Two months ago, during DrupalCon San Francisco, the number was around 115, so since then the number has dropped by 50. Of course, new critical issues has been discovered and added, so the real number of solved issues since then is much higher.
The graph at the Drupal 7 Release Date site gives a good indication about when we can expect RC1 (release candidate). Since my Drupal 7 Projected Release Date August 10 post 16 days ago the new projected date is August 25th. Basically, the projected date has been pushed one day forward every day since then.
That's not really what we want to see, is it?
What can We do to Help?
Right now the best way to help is to dig into the critical issue queue and provide patches to iron out the bugs preventing Drupal 7 from getting ready. However that requires very good PHP and Drupal development skills, which not everyone has.
Other ways you can help is to test patches and report back to the community how they work. The more of you that does that, the quicker these problems will be solved.
Personally I opted for a more unorthodox method, namely launching this website on it as described in my Being Brave with Drupal 7 alpha 5 post.
Since then I have ran into quite a few issues, here is the breakdown of the ones I reported:
- 12 new issues reported
- 2 that already existed that I added information too. One of them actually has existed since v4.7.2...
- 1 that was marked critical
- 2 that was marked as duplicate (not always easy to find using search)
- 1 solved, the critical one
- 2 marked as Needs Work
Several of the issues I found are quite trivial, such as Page title doesn't change in Top Pages Page column and will not prevent Drupal 7 from going beta, RC or even official release. But they need to be reported and fixed at some point.
On top of the above I have also found issues with several of the contributed modules I use here on Nutshell. Many of them was quickly fixed by the maintainers.
When will Drupal 7 be Ready?
Nobody knows, we can still only guesstimate. Hopefully the number of critical issues will start dropping at an higher phase soon.
Pressure from the WordPress 3 release is good, but what's important is that it has to take the time it takes. We don't want to rush out a release that is not ready.
In the meantime I hope you all join in and help with whatever you can to get Drupal 7 ready.
- Log in to post comments
Comments
TallDavid (not verified)
Sat, 06/19/2010 - 23:14
Drupal 7 Release dates
Unfortunately, the release date site mentioned in your article does not give a good indication about when we can expect RC1. The statistics used to calculate the "projected date" are not well done. The site uses a very simple linear regression calculation on data that is not linear. Furthermore, the variability of the data about the mean is not taken into account and no confidence intervals are projected.
One way to see the problem with the current method is to look at the "projected dates" posted so far. Each day the predicted date increases further into the future. If the model was accurate, the predicted date would increase and decrease in a roughly random manner as more data points are collected.
I've done a few calculations and curve fitting with the data and find that models other than linear do a better job of explaining the variability. Both exponential and power curves fit the data better (lower R^2) and both project RC1 dates further in the future. Even though these curves fit the data better than the linear fit, they also do not provide predictions with high confidence.
If the current level of effort and methodology for addressing Critical Issues remain constant, the data predicts zero critical issues most likely in the September-November time frame. However, even this 3 month range is not guaranteed.
If time permits, I'd be interested in exploring time-series and EWMA methods of forecasting the date. If anyone knows of better methods applicable to this problem, I'd be interested in taking a look.
thomas
Sun, 06/20/2010 - 00:04
Re: Drupal 7 Release dates
Hi TallDavid,
Thanks for your comment, and I do agree with you that the website does use a rather simplistic method to calculate the projected release date.
That's one of the reasons I ended my first post with the Mark Twain quote about "Lies, damned lies, and statistics".
This method would be more accurate if the number of critical issues would only go down, but sine new ones are added as well, it makes it more inaccurate.
However, I don't fully agree with that it does Drupal a disservice, as you mention in your comment to the other post. I think it is quite harmless, and also tells a lot about the creativity among people within the Drupal community.
TallDavid (not verified)
Sun, 06/20/2010 - 00:45
Re: Drupal 7 Release dates
Thomas, the method would not "be more accurate if the number of critical issues would only go down". He's using a wrong statistical method for this data set. A good model would work well with the real-world data which includes both increases and decreases in the number of critical issues. Heck, the first prediction from the linear regression model used by @ActivismLabs says that Drupal 7 RC1 will be out tomorrow, June 20, 2010 (see: http://drupal.org/node/770776 ). Care to make a wager on that one?
I think a better model could be developed if the total number of critical issues was broken down into active, needs work, in review, etc. by date. Unfortunately, I don't believe this data has been collected over the past few months.
I've seen many people re-tweeting this bad prediction and I'm afraid that some people are using the bad prediction for planning purposes. Since known bad information is being spread widely and since most people are not sufficiently versed in statistics to recognize the error in the methodology and since some will use the bad dates to make bad decisions, I stand behind my assertion that these predictions are doing a disservice to the Drupal community.
Know any good statisticians that we can call on for advice?
thomas
Sun, 06/20/2010 - 01:26
Re: Drupal 7 Release dates
I'm sure a much better model can be developed. If "projected" was changed to "predicted" it would probably lighten things up a bit.
On the other hand, Drupal is an Open Source community project and the release date is a moving target, which his graph in a way reflects.
TallDavid (not verified)
Sun, 06/20/2010 - 01:47
Drupal 7 Release dates
Thomas,
changing "projected" to "predicted" would make no difference.
If the graph was presented as just data points, without the incorrect fit line, I could see value. The problem is with the bogus predictions/projections/guesses that are represented as being based on good science.