|
There are 11 blog entries within the category of Site Pages
I've probably had hundreds of people over the years request a mobile version of this site. I suspect the reason it's not already mobile-friendly is just one of those things that non-technical people can't fully understand. Given my actual website contains so much content and given that I'm the sole person who maintains it (and who has coded all of it), modifying the overall format is actually a pretty mammoth task.
There are generally two ways one can approach this problem:
1) The entire base website can be rebuilt into a responsive design fashion so that everything adjusts accordingly regardless of the interface accessing it (mobile, tablet, desktop, etc). This is a wonderful design option, but would require an enormous overhaul of the existing layouts. I'd have to go back through every single page and build the page entities to work in a responsive format. In short, this would be a rather significant amount of work, namely as it would require me ensuring that the desktop maintained its design elements too. This is further exacerbated by the reality that I'm a developer and not actually a designer (despite the fact that I do my own design too).
2) The second option is to build a brand new mobile version of the site that runs alongside of the desktop version. The downside is that I'll have two separate sites to maintain, but the upside being that changing one of them doesn't requiring the maintenance of two of them.
For various technical reasons, I've opted to go with option number 2. The bulk of my website is based upon complicated backend code that handles all of the requests from the database. The front-end is generally pretty straightforward and indeed works well on a desktop. It just doesn't work well on a mobile device. Incidentally, the reason it works well on a mobile phone is because the entirety of the site is built exactly to HTML 5 specifications. This is what allows the site to properly "shrink down" when viewed on a mobile device. The downside is that it becomes nearly impossible to read and navigate.
So essentially what will happen is that a new website will exist for mobile users. The site will incorporate the standard mobile-based navigation options, typically known as the "hamburger" dropdown. The backend code will remain in place and will be used to drive the new front-end design. Some backend changes will need to be added, but the majority of it will already be in place. On the other hand, the secondary design elements will need to all be created.
I've been working on this on and off for some time now . It's finally gotten close enough to being finished that I thought it was time to mention it.
As a quick side note, I've spent the entirety of 2022 completing a dozen pretty huge goals that I'd set out for myself. I only started working on them in March and yet, I should be on track to finish every one of them. Some of these goals include significantly elevating my work life, rebuilding part of my house, finishing my camper van, recording an album, rolling out several new software platforms, rolling out a new version of kevinludlow.com, and indeed rolling out the mobile version of kevinludlow.com.
I'm extremely confident at this point that the new site will be available soon enough!
|
A few days after releasing the new blog format, I found there was a small bug with the grouping of older month/year combinations. I've since fixed that code and everything seems to be working as expected now.
I should also mention that there has been a great deal of work that's gone into maintaining the page Sitemap. This isn't intended to be used by any regular person, but rather should help the various bots to better index and catalogue the full extent of the site.
|
I've been meaning to write a post for a couple of weeks now, but I've been especially busy with work!
I've been fortunate enough to be implementing all sorts of new innovations at work and so I've had to balance my development time accordingly. But between the massive code changes to this site (and ultimately the software that runs it), the various publications that I've been working on, the handful of TikTok channels I've started managing, the music work I've been trying to stay on top of, and the numerous travel plans I've been working on (namely to Alaska, Hawaii, Maine, and Africa), things have been extremely busy!
While the site might not look super different just yet, all of the new development code has been merged into the main code branch and I can absolutely say that the changes are plentiful.
One of the biggest changes to come out of the new release has been the addition of my own Blog management software.
How I Originally Modified Wordpress
When I first wrote this software back in 2005 (at which time the project was called staticMOD), I actually had taken the time to write my own blog management software. Blogs hadn't been around all that long and they wouldn't last much longer with the advent of YouTube mostly taking over the space. So although staticMOD relied upon a custom blog interface, the successor codebase. called openFace, did not. Instead I simply modified the backend of Wordpress and used part of it inside of my administrative toolset.
None of my site was actually running on Wordpress. Rather, there was just a nice text editor built into it and so I customized it so that the entries would be saved into my own personal database with my own personal flair. It's worked extremely well for the past 15 years, but I've been wanting to officially move away from it and so I finally have. My goal has been to have as few 3rd party applications as possible.
The Process of Custom Blog Management
I started making these changes in the summer of 2021, but then had to put the project on hold for a little while. I picked it back up in November and made huge strides - generally getting to work on it most every evening. But with many more personal downturns still ahead, I ultimately had to put the project on hold for a few months - which I did. When I finally came back to all of the coding around March of 2022, it's pretty much been a steady sprint to the finish line. And here we are.
There are definitely still some things that I'm working out, but all in all the site is much easier for me to manage. I now have the ability to easily start integrating all of my social media into my own site. I don't think we've yet reached a time whereby people truly value that, but some day they will.
Additional Tech Highlights
There are two additional tech highlights that I'll mention as they've both been incorporated into my software.
In December of 2021, I put a ton of work into a language processing tool. It was essentially designed to be fed any kind of communication between two people and to analyze every little nuance of said communication. I thought it was a pretty cool tool. While I had fun building the project, it ultimately proved to be a huge waste of my time and effort (the project did take me a ridiculous amount of time to make and like most of my projects was coded entirely from scratch).
But as I'm always one to find the silver lining, I took that software and instead retooled it for openFace. I haven't released it just yet, but it has the ability to analyze decades of information that I have and break that information down into much more meaningful pieces, all of which will have accompanying timelines.
And then in April/May of 2022, I finally was able to build my own facial recognition tool. I've only recently found a good script that will allow me to use this, but my hope is that I'll be able to incorporate this into openFace and ultimately have my entire photo (and later video) library automatically catalogued for me with primitive AI.
More Changes Coming
There are still a host of changes that I haven't rolled out yet. As I had previously written, all of the photos and videos are now in a distributed cloud network. They load very fast and much more importantly, I finally have an infinitely scalable filesystem. I will definitely be releasing these assets more and more in the coming weeks and months.
|
 Installing one of my new 3TB drives into the Synology RS-411
In addition to the new disk space has come all sorts of redundancy, which quite frankly is much more important to me. Not that a disk failure wouldn't still be a huge pain in the ass, but at least I am no longer worried about data loss. There would have to be a pretty cataclysmic failure at this point to lose anything. Many thanks to Dave G. for all of his help in picking out equipment and helping to get things setup. Hopefully his Amazon coffers are a bit better padded now. Over the coming weeks I'll be working on solidifying the new site layout and getting everything online that I want to. It will be several more months before I have all of my new content fully displayed, but I'll get there. Hopefully I can start keeping things updated at a much more rapid pace from here on out.
I'll post some pictures of the new setup just as soon as I have some time.
|
It's been a couple of months since I finally decided that it was time for a full overhaul of kevinludlow.com and there have been some pretty significant changes already made to the architecture of the site. I was joking to DaveG just a couple of days ago how significantly my CMS architecture skills have improved since I first developed the concept back in 2005, and even since formally releasing it in 2007.
The size of the codebase has been significantly reduced and puts much more strain on the SQL server for larger, more complex operations (as opposed to parsing data via code after simple queries). Additionally, most all of the original Javascript used has been replaced with jQuery or removed entirely. While I'm thrilled to have done this, it's not entirely an original design flaw. When I first began coding the system, the jQuery library was not yet available and developers were essentially limited to the Prototype / Scriptaculous packages. To this day I contend that those packages are hugely bloated and should not be used for web development projects given the complexities that they add to the maintenance of projects. This was especially the case in 2007 when I first released openFace given processors were that much slower and front-end Javascript weighs a browser down. Fortunately jQuery provides an excellent balance of functionality and programmability while still remaining pretty light (provided one avoids their various UI packages).
In parallel to this development (which is really just in my spare time), I have been hiring people to take on the arduous tasks of scanning photos, digitizing movies, digitizing old audio recordings, and etc. As of this writing I already have 11,254 new photos that will be added to the website once it is fully launched, though I'm guessing it will be closer to 17,000 new photos once I'm finished!
|
In early 2008 I came across an interesting idea that Amazon had started working on called Mechanical Turk. The idea was pretty simple: rather than writing software that was capable of artificially intelligent tasks, it was instead plausible simply to pay people very small sums of money to perform very small tasks. This is what the Turk program did.
The Amazon system allowed developers like myself to create small programs that would integrate with their Turk API. In my case, I wrote a very simple program that would display a random photo from my collection and request that the user describe the photo in the text-box I provided to them.
All of this was spelled out on within my user space, but provided that the person described the photo in at least 10 words, they would be paid a sum of $0.01 (1 penny). I suspect that this type of thing is mostly aimed at people in English-speaking emerging markets around the world (India for example), but proved to be invaluable to me. On the initial 3 month run, I wound up having over 8,000 photos described.
Update March 3rd, 2012: I've done some additional exploring of this system and in the few years since I first started experimenting with it, Amazon has made huge strides. In 2008 their developer API was pretty cumbersome to figure out, but they've really managed to solve this by simplifying the submission process. The downside is that it requires a few extra steps on the developer's end, but the upside is the simplicity to understand how it all works and to begin a campaign. The dataset importation used to be automated, but now one just uploads a CSV file of the data. Again, it's an extra step, but is so simple to do that I can't really complain.
All that being said, I've recently launched another several thousand photos and had 100% of them described within just a few days. For the most part, the entries submitted were very good. I will likely do a full-scale launch within the next few weeks and have Amazon Turk workers detail ALL of my photo collection. This is a pretty mammoth amount of data processing that people will be doing for me, and I'm excited to think how it will change my search criteria.
Update October 28th, 2015 (small political rant within): After many, many years of wonderful successes with the Mechanical Turk project, I decided that it was time to once again overhaul some of my older code. I spent several days rewriting a bunch of the internal mechanics of openFace (the software I designed that powers this entire website). Given the importance of this piece to me, I focused heavily on the Mechanical Turk processing. From a code point of view everything went great. As it happens though, there is now a community of Mechanical Turk "watchdogs" (they have no affiliation with Amazon by the way). This was an interesting experience for me.
The new scripts I wrote (to compliment the detailed classes I've already written) were probably a few thousand photos in when a small group of people (Californians as it happened) started sending me angry emails. In a nutshell, they told me that the price of each photo should be more like $0.30 - $0.40. Keep in mind it takes about 15 seconds to look at a picture and type a 10 word sentence about it. Always one to engage conflict, I tried reasoning with them. I explained that this was just a personal project and that I couldn't possibly spend that kind of money on it. At $0.01 - $0.02 per photo it costs me about $40.00 for 1,000 photos (the fees are over 100% at such a low rate). Still, when I process 8,000 photos it can add up. So while the fees would be a lot less at $0.30 per photo, 1,000 photos would wind up costing me nearly $400.00. It then follows that 8,000 photos would cost me about $3,200. I explained to this group that if it *did* cost that much, I simply wouldn't be able to use the service and then nobody would get my money. It's a hobby of mine to catalog my life and the time I share with those people important to me; it's not a business. It's all at a financial loss.
They weren't having it. They were angry and not shy about letting me know. But then I took a step back from their complaints and I realized that like most marketplaces, the people who are doing the work are extremely delighted to have a steady source of income coming in. Many of those people might live on just a few dollars per month and so the opportunity to make even $20.00 from me is huge to them. For some of them it can entirely change their quality of life and open new doors. Of course their lives are nothing like my own. I'm sad for that and would do most anything to change that. I'd sure as shit rather help a group of people from emerging countries like Brazil, India, China, etc. than have a bunch of people from California complain that I'm not doing enough to help THEM (while they sit in an air-conditioned Starbucks sipping lattes and yelling at me from behind their MacBook Pro).
So like so many things in life, once you tune out the background noise, the Mechanical Turk project continues to be pretty amazing. I'm delighted to give money to these people all around the world; they do a wonderful job for me.
|
|
