A new blogs.perl.org or die trying
TL;DR: 1,600 ranting words including F-bombs abound. I am testing out a new theme for this site. Also I am trying to port some content to a new CMS and put up a public test site to get feedback.
Where to begin... how about the latest problem I had. I tried to post a comment saying I fixed a mistake in my last post and it tells me I am not signed in. After logging out and in then trying it a few times with no success I looked to the github issue list. I find the following issues that seem to match my problem #100, #196, and #222. Seeing that this is a known problem I decide to see if I can fix it. I decide to start by setting up a local copy of Movable Type (MT) to experiment with. So beings my tale of woe.
Is Movable Type actually open source?
I go to movabletype.org and click on "Getting Started" and it takes me to a page where it says I need to purchase a license to "get" it.
Next stop GitHub. I find movabletype/movabletype which appears to be the canonical source. Browsing the Readmes I find no information about the license and there is no documentation directory either. I grep the code for license information and find the following message:
# Movable Type (r) (C) 2001-2013 Six Apart, Ltd. All Rights Reserved. # This code cannot be redistributed without permission from www.sixapart.com. # For more information, consult your Movable Type license.
So nothing about being open source so far. I bust out the search engine and am led to the MT wikipedia page which states MT is under the GPL. I check the citation links to find this OSS version and I get 404s and pages with no content on them.
Dark Pattern: Shit documentation
MT has no documentation that comes with the code and the stuff on their website is piecemeal. I wish I could say this was uncommon but I fucking see it with every project like this. A seemingly useful program that is “open source” and yet to get real support and docs you need to pay a company for it. I started to get this feeling as soon as I saw how flimsy their Readme.md was. I find this practice just revolting and I have never recommended a program that employed this technique to get money out of the users.
MT's default settings
You know what makes an application great? Sane defaults and MT does not have that. First lets look at saving an entry. If I start work on a post and click the save button by default it publishes it. Who thought it was a good idea to have a content management system (CMS) default to publishing entries upon clicking save? I thought hit save meant just saving the damn thing. How many times have people created an entry, saved it, and then realized their half formed ideas were now public? That does not sound sane at all.
Next is the date handling for when an article is published. When you create an entry it populates the “Publish Date” fields with your local date and time. Local date and time... timezones anyone? You see the results of this idiotic design decision on the front page of the site all the time, hell right now it is happening. My last post was at 2:42 AM local time, burnersk did a post at 9:40 AM his time and if I were to do another post that was before 9:40 AM my time the front page would still show his at the top instead of the actual newest post made.
That shit is insane. I could game the publish time to guarantee my posts stay on the front page of the site longer and in a more prominent position, above the fold. This is a terrible default setting.
More popular CMS products do not allow normal users to edit publish times by default. This feature is usually reserved for admin users. Another way to set it up is to allow the users to set a publish window and have it approved by an editor before going live. Online news sites use this technique to make sure they always have new content coming up to feed the 24/7 news cycle.
Oh there are so many UX issues, lets just roll with one anyone who has used MT knows about. Body vs Extended. WTF is this? Body means body of the article right? And extended means footnotes and the like? Wrong. Body really means article teaser or abstract and Extended means the article body. MT stacks the Body on top of the Extended in the full article view so you have to split off the initial content into the Body. This is just dumb and the terminology does not help. Issue #194 is about educating users on how to properly use MT's extended feature. Instead why not have the CMS generate the abstract based on a set number of words? Because that would be like all the other CMSes and we cannot have that. Looking at the definitions in the Creating and Publishing Entries only makes me bang my head against my desk. Who the fuck do they think their audience is exactly?
Tags and Keywords. Two fields that can mean the same thing in certain contexts. The docs say tags are for tagging and keywords are to insert keyword metadata into the page's header and can be displayed and used for navigation. So tags + metadata = keywords. We don't need both. The audience who uses this site is interested in posting content and that is it. All these other features are just getting in the way. The default create an entry page could easily be stripped down feature wise and it would be a better fit for the majority of users. I am not saying get rid of features all together just don't show most of them by default.
I almost forgot there are categories too.
Adding assets is inconsistent. If you add an asset via the list on the bottom right you get the “Insert Asset” page and then upload the image. Then you click “Insert Image” in the WYSIWYG, select the image and click continue. You fill in some File Options and click finish. You now have a generated bit of html for displaying your image. If you start the process by clicking “Insert Image” and then uploading an image when you click continue you get the File Options screen. A File Options screen with two more editable fields that were not present when the File Options screen was displayed in the first workflow.
Rage based development
I had been tossing around the idea for a new theme for a while. I contacted the site maintainers and back in July Dave Cross emailed me back and said there was interest in these ideas. Since then I have been playing around with different pieces of the site to build a quality look and feel. This latest series of annoyances with MT caused me to dig up a blogs.perl.org scraper I had written at the end of 2011.
I was so pissed off then about MT that I was going to try something else. I forget who it was but someone suggested I just didn't have the hang of MT yet and I needed to use it more. Once my comfort level increased I would find MT usable. So I set aside my plan to try created a new site. 27 posts and a slew of comments later I still think MT sucks and here I am again.
A new theme
This theme I have is missing one key ingredient before I unleash it unto all of you for review and comment. Testing it on Apple products or rather fuck Apple. Testing a website on Firefox, Chromium, or Opera/Opera mobile is easy. Microsoft provides free virtual machines for different versions of IE. The Android SDK is free and runs on Windows, Linux, Mac OSX. Apple provides nothing outside of their ecosystem. You have to run the iOS emulator on Mac OSX. Apple is the only major player who I would have to buy some piece of their equipment in order to test for their other equipment.
I don't fucking think so. Normally when I do website design for clients they can provide an Apple product for testing. I make sure to mention this upfront and it never turns out to be a problem. I currently have no prolonged access to an Apple desktop to do testing. This is where I am at with the theme. The online iPhone/iPad emulators are junk, it is just a cut down frame, not an actual native environment. I am working on getting access to an Apple desktop but I have no ETA on that or even if I can. Most of the people I know use either Windows or Linux.
A new CMS
My plan for this is pretty straight forward. Use my scraper to grab a series of articles and profile pages. Take that sample data and input it into another CMS. Setup said CMS with the new theme, test, and put a beta up for public consumption.
The next post I put up on this site will say one of two things. The ideal post will be announcing the beta CMS is up online and ready for people to try parts of it out. The less ideal post will have the new theme setup as a static version of the current site for people to try out. Either way there ain't no gettin' offa this train we on!