Are We Hurting Our Clients, Business and Site Visitors by Supporting Crappy Hosting?

Comment by Joe LeBlanc on May 24, 2010 at 3:36pm

This is a bit of a tough call for me. I agree that if you're in the business of building sites, taking a hard stance on getting good hosting is going to be best in the long run. I remember one client of mine who asked if their usual host [unnamed] was good and I simply said 'no, here's the one we should use.' They didn't put up any sort of fight and the site launched successfully on my suggested host.

However, if you're creating a reusable extension in the PHP world, you're inevitably going to run into inconsistencies. This is bound to happen even when hosts are running the latest and greatest version of PHP. Although I try to push for better baseline setups, the line is a little bit fuzzier. For instance, when I upgraded Podcast Suite for Joomla 1.5, I said "no more PHP4, it's all PHP5 now." I proceeded to use a PHP5 XML library to do all of my RSS generation. Lo and behold, some configurations are set up to disable this standard XML library. I decided not to rewrite Podcast Suite to get around these issues, mainly because Podcast Suite is not a money maker for me.

In Nicolas' case, it's highly likely that many of his customers use AkeebaBackup to help themselves get off crappy hosting. Even though it takes him more time to produce and support a broader range of hosts, it may very well be worth it for the software that AkeebaBackup is. I doubt Nicolas can effectively offer months of free hosting to all of his customers that run into issues, so it may end up being more cost effective for him to simply work around the oddities.

So yeah, I think we need to push for high standards when we can, but the closer you get to the code, the more you have to pick and choose your battles.

Comment by Robert Vining on May 24, 2010 at 3:46pm

All good points Joe, and ones I don't consider as 'just' a site developer and not an extension developer.

And I need to say here too, this was not to drag Akeeba Backup as the topic of discussion, far from it. I believe Nicholas is doing great things with the products he offers and go as far as to support them with a link from my own home page. So I don't mean to offend him with my discussion.

I hope I'm clear on that and Nicholas, contact me and I will gladly rewrite this post without your quote if I have taken too much liberty here in quoting your response from the other thread!

It just sparked a thought that I have been pondering for quite some time, and I hadn't fully acted on with my own clients until a few months ago. But I'm glad I'm in a position to do so and have see positive results because of it.

Thanks for your insight Joe. Things to consider from the 'other side of the fence' for sure.

Comment by Joe LeBlanc on May 24, 2010 at 4:42pm

Your post is fine and this is a good discussion to bring up. I didn't really take it as a pass at Nicolas, but feel that Akeeba is a good example of where the dilemma in all this lies.

I can feel where you're coming from and understand the frustration. Joomla's core strength and weakness has been striking a balance between: "whoo, it's installed and working!" and getting it to install on as many hosts as possible. This unfortunately places us squarely in the middle of the cheap hosting dilemma, while it does make us very versatile.

Many in the PHP community at large tend to look down their nose at Joomla and WordPress because we tolerate shared hosting setups. While I think this is uncalled for and a tad snooty, I also realize that we have a responsibility to raise the bar when possible.

Comment by Bo Astrup on May 25, 2010 at 12:08am

At the end of the day it comes down to education and knowledge - and "fighting" great marketing (read GoDaddy (as an example) ) In Danish we have a saying: "just because there are 1,000 flies sitting on it, doesn't mean it's any good) More often than I like I hear that GD word when someone asks for hosting.

Now, to me one issue is how about a white paper on the "dream" host configuration for Joomla! ..? That being said I know this will turn some people off as it's will most likely be called favoritism. BUT in real life I have run into quite a few specialized (small) hosting companies who would love to be a good Joomla! host but do not know what is required. As a matter of fact the most issues I have had with host was a small two man hosting company and it was not worth my time in a situation where it was more an issue trying to convince this individual to change apache settings to make it work.

Again I believe it would have been helpful to have a whitepaper that would outline the ideal hosting setup. I realize we can not have an end all be all, but having some kind of a guideline would be helpful.
Or cheat sheet for the client to use when shopping for hosting.

Comment by Hannes Papenberg on May 25, 2010 at 3:26am

My stand on this, especially as someone who has worked on on the Joomla core extensively, is, that we should punish bad hosts.

If I look at Joomla 1.5, 1.0 or Mambo, we've always kept back because this or that feature isn't available on really every server and we thought that we had to support each and every server setup, regardless of how ancient it is. Likewise with browsers. Joomla 1.5 would be 3 to 4 times as fast if we had used PHP5 for this. Now we have 1.6 with a huge amount of new features and an insane flexibility and we are still two times as fast as 1.5. Sticking to supporting crappy hosts means, that we are not only coding more than necessary and have to give to much support, but it also means that high traffic sites need more servers to handle their website and hosting companies can have less low-traffic sites on one server, which hurts both the environment and the economy.

We are at a similar point with browsers: Supporting IE6 or even IE7 for the backend means a lot of jumping through hoops and compromises and it also means more load for the hosts again. One example that might look tiny, but accumulated over 4 million websites makes a difference: The current Joomla backend template has those nice rounded corners. These are made out of 6 additional divs and 4 images. If we would use CSS3 for this feature, we would save these 6 divs on each page load. This would mean several gigabytes less of data to transfer between server and browser when done everywhere.

The thing is, that Joomla is not a small project without influence. We most likely come directly after wordpress when it comes down to deployed installations and when it comes to the total economic value, I'd even go as far as saying that Joomla is the market leader here. All those webshops, company websites, etc... We are creating standards, we are the ones that can push hosting companies and users to move to new, better systems. But if we stick to the current crap and keep on still supporting it, that will never happen. As long as everyone still tries to support IE6 and PHP4, we are going to still see that crap around. There are no reasons why you should still use IE6 or less and there are no reasons why you have to use PHP4. (I know that there are companys that have solutions based on IE6 and that there is software out there that needs PHP4, but do you also accept working in a building with asbestos and lead paint just because your boss is to lazy to move to another building?)

Please let us make a move and kill both PHP4, IE6 and hosts that think they have to define their own version of PHP.

Comment by HarryB on May 25, 2010 at 6:33am

Good point Brian...and ten years from now, somebody will be saying "It's not a cheap decision to move from IE7 on 20,000 Win XP machines."

Unfortunately, in the real world Joomla! (nor Wordpress, nor Drupal, nor ...) does not really have the power and influence to "make a move and kill both PHP4, IE6 and hosts that think they have to define their own version of PHP." PHP4 and the majority of bad hosts will probably disappear into the noise over time as SaaS takes hold and drives out smaller players, but killing IE6 ( or IE7) anytime in the foreseeable future is, in my opinion, "Mission Impossible."

You may disagree with the above, but,in my opinion, truth be told, it's a fact of life.

Comment by Hannes Papenberg on May 25, 2010 at 9:19am

Yes, of course it is not cheap to update 20,000 win2k machines to WinXP or Vista or 7 to update IE to something better than 6. At the same time you could install FF and be done with it. If you have difficulties convincing them that IE6 is the devil, just point to all those security issues that it has and show them how much productivity they are loosing because of this piece of crap. I know that you have to get a little bit more creative than that. anyway...

I still say that unless we make a first move, the others wont move. As I wrote before, it is not just a question of "Do we loose money if we don't support it?", but it is a question of "How much money are we going to loose if we still support it?" How much more traffic do I have to pay, how much more development time do I have to invest, how much more server power do I need to provide, how much more support do I have to give because of this program? The longer I think about it, the more I wish there was a law banning this program.

If we start now to only support browsers starting with IE7, it will make it easier for other projects to abandon older browsers and big projects like youtube and such have an easier time to also switch. And you might think that Joomla is not that powerfull, but I can guarantee you, that our influence is much bigger than we think it is.

Oh, and I expect the number of IE browsers to go down pretty drastically in the future when all new windows installations in the EU get a window to choose which browser the user wants to install. Sometimes I love the EU... :-)

Comment by HarryB on May 25, 2010 at 9:58am

"..I still say that unless we make a first move, the others wont move...." Join the crowd. ;-)

Joomla! may well be a world power, but in most circles I travel in, mention of the word "Joomla!" more often than not draws a blank stare! ;-)

Regarding the EU...they should give thanks daily that their Microsoft Cash Cow exists!

I'm done now... more important stuff to do...

Comment