Australian government likely to standardise on Drupal
computerworld.com.au
computerworld.com.au
Meanwhile, they had a grand opening at Parliament House to which all the media were invited -- and I was very pointedly not invited, presumably because they were worried I'd say rude things about them (spoiler: I would have). Then the leader of the National Party chose that morning to resign, and nobody from the media came to the launch party because they were busy over the other side of the House interviewing him. So I felt a certain degree of whatever that German word is for "serves you right, you bastards".
Meanwhile, australia.gov.au had a bunch of redesigns and is even better now, so the three guys they brought in to replace me obviously did good work...
Schadenfreude.
They do a lot of research, produce a lot of reports, publish a lot of "standards", and then everyone else just ignores them and stumbles on with whatever they were doing before.
The people who are already married to some solution will remain married to their solution; they will successfully argue for exemptions.
Those who are determined to use something else will use political capital to get another exemption.
Those who are forced to make the switch will make the switch, but the old system will probably hang around for a long time, more than doubling the overall cost due to double-entry and errors.
Some will make the transition, it will fail, then transition back.
In the end the government standard will be taken up by a handful of sites, AGIMO or its heirs and successors will declare total success, the Minister will have been rotated to something else or kicked out at an election and nobody will give two hoots.
Welcome to public sector computing! It's like enterprise, except even crazier.
"GovCMS is intended to support more effective web channel delivery functions within Government, and enable agencies to redirect effort from non-core transactional activities, towards higher-value activities that are more aligned with core agency missions," a draft statement of requirements issued by AGIMO states.
'''
It appears buzzword abuse is becoming an international problem across multiple industries and agencies. I recommend a UN task force to look into it and take the appropriate measures. Sanctions are defiantly in order, but at this point nothing should be off the table.
However, there used to be more U.S. Drupal sites, including Data.gov...there was even some hubabaloo about them releasing the source code:
http://www.whitehouse.gov/blog/2011/12/05/datagov-goes-globa...
https://groups.drupal.org/node/194808
Viewing the source code of the current data.gov (view-source:http://www.data.gov/), it looks like the moved it to WordPress. That's a pretty damning migration.
Edit:
The current repo for Data.gov can be found on Github - https://github.com/GSA/data.gov/
Here's the shit-show that was apparently Data.gov's Drupal installation: https://github.com/opengovplatform/ogpl-d7
And quite humorously, it seems to be something that we're exporting to other governments, such as India https://github.com/opengovplatform/opengovplatform-beta
From CKAN's webpage:
> Strong integration with third-party CMS’s like Drupal and WordPress
There was no explicit reason to switch front-end sites...But think about it...creating Data.gov was a large endeavor, but it's certainly not a mission critical one. Just a couple years after building that initiative from scratch, they throw out Drupal for WordPress instead of using their existing Drupal installation and installing CKAN modules. How does that not reflect badly on Drupal, which bills itself as a much more powerful, flexible CMS than your standard blog system?
I used Drupal back in the 5.x days and thought it was pretty swell, but that was when I was a rookie developer who had almost no real full-stack dev experience. Now I can't think of a single use case in which Drupal's purported feature set is needed that would not be better fulfilled by just going to Django/Rails. The hackery that seemingly has to be done to meet clients' perceived needs almost eliminates whatever advantage Drupal in terms of a community of knowledgable developers to hire from. And I would question whether there are really more Drupal-only developers than Rails-only developers (certainly, there are way more Wordpress developers).
I literally will not touch Drupal projects, I'll go back to working retail before I'd be a Drupal developer again, it is an unremittingly awful platform to work on.
The front end experience is horrible (even with all the tweaks in the world), CCK/Views made it just about useful then they rolled that into core but only some of it, you can forget an upgrade path between major versions in any real world scenario, the hook system is just plain bonkers.
I honestly don't know how it got a reputation as the goto open source enterprise CMS, Joomla must have been truly terrible for Drupal to be the preferred solution.
In the time I spent beating that monstrosity into shape I could have built the same functionality in a fraction of the time using something like Symfony.
This was in the days of the 6 to 7 rollover, things may have improved, frankly I've very carefully not looked.
There are also plenty of bots that scan for vulnerable Joomla installs as they do with Wordpress.
The 'bonkers' hook system is basically just aspect-oriented programming. It's not to everyone's taste but it's not as though you can't figure out how it works. It's just not a typical MVC web framework, and MVC web frameworks have mostly won (which is why Drupal 8 is going to be built around Symfony components, but that's another story).
I dispute that an experienced Symfony dev could build the same functionality as an experienced Drupal dev in less time. You could skew the test either way by picking requirements that favour one platform over the other, but for most general case content management (including user-generated content) requirements you'll be able to do it quicker in Drupal, mostly because you won't have to write boring stuff like:
* User management
* Roles and permissions
* Login, password reset, activation
* Basic editorial UI
* RSS feeds
* Custom content types
* Wysiwyg editing (admittedly this can be a pain to configure)
* Logging
* Caching, including Varnish integration, Memcache etc.
* Indexing of content into Solr/Elasticsearch
And so on. Most of that stuff is plug and play, and if you're happy to accept the (very real and sometimes painful) limits that come with Drupal then you can almost always deliver functionality faster by using open source modules wisely.
In general, I rarely agree with people who say "<popular piece of software/programming language/library> is an unusable piece of crap" because the evidence appears to suggest that other people are using it effectively. The fact that you didn't enjoy it doesn't automatically make it bad software (nor does it make you a bad developer for not enjoying it or figuring out how to use it effectively). Most successful systems do at least some things really well, even if they get a lot of other things badly wrong, and dismissing them makes it harder to learn from the things that they do well.
For my part, I wouldn't exactly describe Drupal development as fun, compared to Ruby or Clojure. However, it can be scarily effective in the right hands.
I've built custom CMS's for clients that where a much closer fit with their existing workflows that implemented most of the list above (not everyone requires everything above) and I've done it at least as quickly as doing it from scratch in Drupal would have (partially because I have existing code I can reuse for a lot of that).
My visceral dislike of Drupal comes from working with it directly for two years, I read the books, I went through the code, I wrote custom modules, I learnt CCK/Views inside and out and at the end of it I took a pay cut to not have to work with Drupal.
My opinion of Drupal can best be summed up with an analogy - given sufficient thrust even a brick will fly and Drupal required a vast amount of thrust.
Yes out the box you get a great deal of functionality which is what makes it so enticing but then your needs change and you require doing something that is outside the scope of Drupal or even existing modules and the whole thing comes crashing down.
Back when I worked with Drupal it created hundreds of tables for anything involving slightly complicated CCK/View based workflows, performance was horrible for logged in users (and trust me I haunted 2bits like a poltergeist for every tuning opportunity I could find).
But then this is just opinion at the end of the day.
One thing I will say is that the Drupal community are an absolutely lovely bunch and incredibly helpful which is always a strong plus.
Also just for humour value http://www.cmscritic.com/wp-content/uploads/2010/08/permissi... this is not a good UI.
The draft Statement of Requirements for GovCMS: http://www.finance.gov.au/sites/default/files/GovCMS%20State...
Some notable tidbits:
- No analysis seems to have been done on the approach of hiring their own developers to develop a custom-yet-cohesive solution, a la Gov.uk.
- The CTO says "Drupal appears to offer the best enterprise-level as shown by our previous extensive analysis". Can someone refer me to this extensive analysis? Have they considered the security risks of standardizing on a single platform? http://www.cvedetails.com/vulnerability-list.php?vendor_id=1...
- "A solution hosted offshore would not be excluded from consideration" So, putting Australian Government infrastructure in the US is OK. Subject to US laws. Fantastic.
- Does anyone know if the government intends to contribute to Drupal development? They say that they expect some "feature rich" sites to require custom modules, but what about Drupal itself?
This is moronic. Why would you mandate using the most unusable CMS system from an end-user point of view when you are outsourcing the hosting and maintenance to a SAAS partner? I'm sure they could have cut a better deal with Sitecore which seems to power a lot of Government sites and is a far more coherent platform with a robust workflow and integration with Active Directory services.
[1] Wordpress isn't really a CMS, but it is treated like one at many corporations.
https://github.com/WhiteHouse/fortyfour
http://whitehouse.gov/blog/2010/04/21/whitehousegov-releases...
1. Perceived as overly complex.
2. Reliant on dev paradigms that originate from the PHP4 days, i.e. no OOP.
3. Performance can be a problem.
4. It's PHP, and old-school PHP at that.
5. Everything is done in the UI, and you can't ship DB-based config in code.
6. It's for people who don't know how to code. (I would argue this applies more to WP than Drupal)
I get all of those points, and I agree with most of them. However, my shop uses Drupal for mid-size projects, and a little bit of WP for small ones. Here's how I see Drupal:
1. Integrated, relatively coherent API and model structure that modules can hook into. I.e. two modules don't need to explicitly know about each other for them to be used together. Thus, uncoordinated effort on disparate modules can be used together by site builders to create novel features.
2. Configuration is much less DB oriented than it was, and many things can be committed to git now. This is also continuing to improve over time, and Drupal 8 will be a major change in this regard.
3. Drupal when properly set up actually seems to provide a more consistent admin experience for users than WP does (in my experience, anyway). WP sites that use a lot of modules usually end up with the admin feeling like it was designed by many different people. Drupal used to be that way, but has greatly improved...
4. Performance with proper caching is actually pretty good. As always, you can use any good tool poorly and get poor results.
5. Drupal's codebase is extremely complex, and structured in a way that makes things very hard to debug sometimes (i.e. the hook system). Sometimes I want to do something that I consider very simple, and I can't do it without digging through the API for 4 hours trying to figure out how to do it.
6. And yes, PHP has problems. I don't like its API and that it tends to encourage poor coding practices, but as some recent threads on HN have pointed out, modern frameworks like Symfony are helping to put a cleaner face on PHP. And D8 uses Symfony.
It seems like many of those that hate Drupal stopped using it in version 6, which definitely had many of the problems listed above, many of which have been addressed to some degree. Drupal 8 is based around Symfony components, which should greatly modernize the codebase, though that has been a very controversial move in the community.
All of that said, Drupal still feels like the "least worst thing" for what we use it for, and I'd love to hear what people use instead of it for the types of projects we use Drupal for (custom e-commerce, large content sites with lots of different types of content, complex publishing workflows, etc). We've tried Joomla and WP at various points, and each felt like it had shortcomings for our use cases. And we don't usually get to work with huge budgets, so we have to deliver a lot of complexity for not a lot of $.
Sorry, this got really long, and this thread seems a strange place to go into so much detail :). But I am genuinely interested in alternatives, and in what others are using when they just need to build a (complex) site for a customer who doesn't have the $ to start totally from scratch.
I gather D7 has improved on this a bit, though it's still nowhere near Wordpress in this regard.
The main trouble I had with the culture was that complexity was seen as something necessary for power; as in, you can't achieve a result with a simple system. You can only chop down a tree with a Swiss Army Knife; an axe is too simple. This carried over into the disastrous decision to throw out all the code, including third-party extensions, in every major version update, meaning that stuff you wrote and tricks you learned for version N would be almost completely useless for version N+1. I found that idea so totally beyond hopeless that I was forced to throw out what I had spent years laboriously learning and move over to Wordpress, which suffers from none of these problems.
So what happened to Squiz? It was the godawful piece of shit that I criticised in the /. post that kinda-sorta lost me my job with AGIMO, so I wouldn't mind knowing how long it took for me to be proved right.
At least AFAIK, Drupal is fairly efficient. I would not be so enthusiastic about Wordpress.
As for Wordpress's alleged speed problems, I haven't had any trouble. The speed of development for WP is way ahead of that for Drupal, in my experience at least, to the point where I just refuse to work on Drupal installations and convert them over to WP as a first step.
It does not support caching out of the box and MySQL has, for a long time, had poor behaviour for workloads involving tables with TEXT fields.
By the way, learn about db-error.php. Might as well never have a problem in style.