The platform for SourceAFIS blog was a hard choice. There’s no satisfactory solution on the market. I ended up putting a self-managed WordPress on the free SourceForge project web servers. It’s the lesser evil of all the other alternatives. It took me quite some time to evaluate all the alternatives and I thought I would write down some notes. I know that thousands of posts exist on this topic, but I have never seen anyone explaining why the CMS market looks like it does today. What’s wrong with CMS?
SourceAFIS previously used SourceForge-managed wiki for all its publishing needs, but SourceForge is terminating its hosted apps offering including wiki. Nevertheless, even if SourceForge continued to manage their wiki app, SourceAFIS would have to move on anyway. Wiki simply isn’t a good publishing platform. It puts nofollow attribute on every external link. It has nearly useless sidebar. It’s authoring-focused where the vast majority of visitors are reading-focused. It encourage poor style that’s hard to read. It’s not really social on small-scale projects. It looks cheap and unprofessional. Fortunately SourceForge also provides managed WordPress. Unfortunately SourceForge-managed WordPress is going to be discontinued this fall along with wiki and all other hosted apps. So what to do now?
SourceForge offers project web where I can install arbitrary applications. But that means self-managed CMS that eats a lot of my time and it is likely to be of poor quality anyway due to my beginner-level knowledge and my limited patience. I have looked around for something better. I needed managed WordPress hosting. Providers for this service are abundant, but they are surprisingly expensive. I would pay $5-$10 annually for this kind of service, but not $5-$10 monthly. That’s hard to justify against all my other expenses. I wonder why these fully automated online services cannot live off a small annual bill? Domain registrars survive on $2 annual margin by automating everything and by being number one choice when users consider premium services. WordPress hosting companies could learn from Backblaze how to optimize hosting for particular application. But perhaps that’s the problem. WordPress doesn’t allow specialization. WordPress is all about plugins and plugins are executable programs that require general-purpose computer to run. Worse yet, WordPress builds on PHP and MySQL, two notoriously inefficient platforms. WordPress was designed to be expensive. New versions and fancy plugins cannot fix it. Inefficiency is at the heart of WordPress design. But perhaps there are alternatives that do not have this flaw?
I’ve considered many other alternatives. There’s an alpha quality plugin that extracts static website from localhost WordPress installation. There are CMS systems specifically designed for off-line authoring and subsequent upload. Compiler, text editor, and special syntax? That’s not even 90′s. That’s 80′s museum artifact. I’ve even encountered some supposedly free website hosting service that threw sneaky upgrade offer at me when I tried to remove its self-promotion link from my website. Ultimately, I believe my expectation of $5/annum or even free blog hosting is rational. It’s just that state of the art has not yet exhausted the potential of the field. WordPress and similar CMS packages are actually heading in the wrong direction. What’s really needed is a fully featured version of Blogger with limited server-side scripting for people who really need it. Until something like that appears I am stuck with self-managed WordPress on sourceforge servers. Possibly later upgrading to either self-hosted or managed WordPress.