Since I had a free day available I decided to take a look at the first alpha release for Drupal 7. I didn't plan on doing anything fancy like upgrading one of my existing sites, I just wanted to explore the features that come with a new site. The experience proved to be a bit disappointing, hence this post.
The first issue I had was mostly my fault, but it was made worse by what happened as I tried to install Drupal 7. I have installed Drupal many times before, both 5 and 6, so I was pretty confident that I knew everything I needed to know about the process. Still, I read and followed the instructions that come with Drupal 7 just so I wouldn't feel like a fool when something didn't work because I wasn't following the instructions. Unfortunately, when I went to the base URL for my site, I got a white screen instead of the installation screens. Over the next hour or so I grew more frustrated as I searched for any mention of the same error and wondered how I could possibly mess up the simplest and most straightforward instructions I've ever seen. Then I decided to append install.php to my base URL and, sure enough, I got the install screens. It was at this point that I learned about the new requirement of the PDO extension for PHP. Now that I knew what to look for I could find this mentioned in the online requirements. (although it's a bit buried) Once I fixed that the installation went smoothly.
Now that I had Drupal 7 installed I first decided to look at what modules are enabled with the standard installation. I was dismayed to find myself waiting 5 minutes as the desired page loaded. It quickly became apparent that the admin pages on a clean installation of Drupal 7 were taking much longer to load than the admin pages of even my most heavily module laden site. The help page took the cake with a 10 minute load time. Fortunately the culprit was easily identified and, after disabling the Overlay module, I got the speedy page loads I expected from a clean site.
If I had the power to do anything to address the issues I faced I would do two things. First, I would add a line to the upgrade notes stating that the PHP extension requirements have changed and those requirements need to be reviewed at such and such a page. Second, I would remove the Overlay module from the standard installation. (Actually, I would remove it from core entirely based on performance issues. That's not a new discussion. #659488: Properly test the overlay to determine if it belongs in core or contrib)
It's still too early to put out any sort of verdict on Drupal 7, but I can say one thing right now. When I looked at the early releases of Drupal 6 I was excited to upgrade my existing Drupal 5 site. With the approach of Drupal 7 I'm feeling mostly dread. I'll test the upgrade path when the release candidate comes out and make my judgment then, when the issues I run into will have real consequences. I could just be feeling shock from the massive extent of the changes coming up.
Comments
Web Browser?
The blank front page is definitely not a good way to greet users upon first load and this behavior ought in my opinion be filed as a bug. Have you done so already?
What web browser are you using? I have never seen longer load times than about 15 seconds on my test sites, most of that time seems to be due to JavaScript doing stuff (and on first load due to lack of cached content). Have you tried it in (Chrome and Safari have excellent JavaScript handling) to verify that you still have a very long load time? What does FireBug or Web Inspector tell you about the load time?
Regarding the overlay, do you have any data to support that the old user interface should be kept? Remember, the overlay is a result of the problems identified during research (by UoB, UoM, and others) that has been published here on drupal.org among other sites (see links in the issue you linked). After reading the linked (very long) issue on moving overlay to contrib or not I am now more convinced that it should stay in core.
A straight line can be drawn from the research to the solution. In the words of Dries: "We know for a fact that not having overlays confuses the hell out of new people (watch almost all of the usability testing videos), and the overlays were specifically designed to avoid that.". The solution needs testing, of course, but that is not an excuse to bash it without providing a foundation for why the solution is bad when the solution rests on multiple test by multiple parties.
Well aware that that the overlay is not the end-all, final destination, or solution for all our administration needs – I am just saying that no real issues have been identified why it is not an improvement over the legacy UI (which is a mess) for the reasons previously stated. I have used Drupal since version 4 and can still stop and think "now where did they put that again?". Not so with the overlay. I have worked as a designer for years now and have a background in cognitive science, that alone makes me trust that these people know what they are doing and what they are doing is the right thing.
I think your dread is valid and you are not alone, we see it everyday with users refusing to upgrade Windows 2000 fearing that their workflow might break or that their widgets they have come to rely upon will not be there anymore. Such is the life of all upgrades. There is a strength in that – people want to know what to expect, but if nothing changes nothing will become better. We have to tread new ground to go forward since standing still is not an option.
By the way, I am having major (sometimes the site even refuses to load) issues with cacheGet() in the latest Drupal 7 releases, but I still think cache should be in core. ;) Sorry about the ranty nature of this post, I kind of felt like I partially had to reply to some posters in that linked issue and not solely to you.
I wanted to sit on reporting
I wanted to sit on reporting my installation issue for a night since there was a certain amount of user error involved on my part. Basically just giving myself some time to think about how it applies to the bigger picture beyond just me.
As for the overlay stuff, I use Firefox. I did try using Chrome and the loading time was much better. There where other issues, but that could be a result of the version of Chrome I have. Back to Firefox, I used FireBug to monitor the activity as I loaded the help page. It took 589 seconds to load and except for a split second at the end where it received data and loaded content from the cache the entire time period was dominated by two activities. The overlay indicated it was connected and the loading circle was painted.
I won't hesitate to confess that I am fairly ignorant when it comes to the topics of usability. My background is meteorology so I haven't paid much attention to those discussions, feeling that they are outside my knowledge base. The only thing I can say regarding usability is what my experience is. I am providing one data point to be considered among many. My suggestion to remove the overlay from the standard installation or from core entirely was based solely on my experience and my impression that it's not ready yet. I linked to the issue to acknowledge there is a larger context than my post. I should state my actual stance on overlay. It's not a bad idea, in fact it's likely a good idea. It just doesn't measure up to what I expect from a core module yet.
The dread I feel isn't due to the fact that this is an upgrade, I jumped on the last one with glee. I'm more feeling dread because I'm getting the impression that this upgrade isn't for me. I would have preferred to see this upgrade spread out a little more. (Compared to the 5 to 6 upgrade, this one is a monster.) Moving too fast is just as bad as standing still sometimes, but I'm not saying that's the case here. It's just a feeling I have.
All ranting is excused when it contributes to a discussion between two viewpoints. If I get another free day sometime I'm going to have to check out those videos simply to get an understanding of what's going on. After all, I can't really discuss what I don't know.
Upgrading Windows is bad example
A better model is upgrading Mac OS X. Upgrade after upgrade hardly a problem. Just sayin'
Drupal install especially D7
Drupal install especially D7 takes a long time to install. Wordpress takes what 5 seconds ? Ridiculous I know :) There's also the lag time when you have many modules in the modules directory, Drupal scans every module and the page loads take longer. With Wordpress plugins you could have a gazillion plugins and the page will still load superfast. Obviously Drupal is doing something very wrong. I think D7 will be an improvement but only a little.
Drupal and Wordpress are two
Drupal and Wordpress are two different beasts with different goals. You're comparing apples and oranges.
Still, page loading time is a legitimate concern and it can be addressed. You should look into static page caching. I use the Boost module to ensure that my visitors have a minimal wait time by loading the pages before they want them.
Turns out my first issue
Turns out my first issue isn't a new one. I reopened #299308: Installing Drupal by visiting index.php (rather than install.php) leads to a fatal error when PDO is not enabled and explained why it isn't fixed yet.