Tag Archives: artificial intelligence

Got WordPress Woes?

Im WordPress nichts Neues.*

Scuttlebutt has it that horrible things are happening (or are about to happen) to the functionality of the Blog Host We All So Admire. This should be no news to those of us who have been blogging here for the past 5 years (as I have), or even longer.

Some bloggers are particularly irate about what they have experienced as a discoverability problem associated with the WordPress search engine. People who commented at one such blog also said they were having trouble finding the blogger in question. The universal conclusion was “a broken WordPress search engine.”

Many bloggers may not have much knowledge about the intricacies of computer technology because they’re “users” (people who are not, and never have been, “programmers,” no matter how many years they may have had computers in their work or personal lives). They never even dabble in HTML, and are content to just plug-and-play. That’s fine and dandy. But not having any background in the anatomy and physiology of computers can make life frustrating when things on the screen seem to be going pear-shaped.

What follows (in the shaded, bold italic block quote) is a short course in computer technology, composed by a person who was programming computers long before the last two generations of geeks were gleams in their granddaddies’ and daddies’ eyes. (It’s highly generalized, but fundamentally correct.)

Computers are stupid. Despite the bells and whistles, all they really understand are only “on” and “off.” What is typed that shows up on a screen as a word is just an electrical signal generated by a combination of alpha-numeric characters that are coded in a computer language. When people talk to computers, they have to use the same language, and user input in that language has to mathematically add up, for communication to occur (hence the name, “computer”).

When you use a search engine, the stupid computer on the receiving end of your request simply looks for exactly what you typed into the search bar. If you leave anything out, or mistakenly add something that isn’t part of the term for which you’re looking, the on-and-off sum that the computer comes up with won’t match anything it can find, and you won’t get any hits. This is not exactly “garbage in, garbage out” (GIGO), because no garbage comes back, although the artificial intelligence (AI) that’s programmed into some search engines will guess that there’s a misspelling, and will come back with some alternative suggestions (which may, indeed, be garbage).

In the recent “discoverability” example cited above, the commentator who said that the blogger couldn’t be found reported having entered as search terms only part of the blogger’s name, only part of the blog’s domain name, and only one iteration of the word that’s used for the blog’s title. These were technically misspellings, but as far as WordPress’s stupid computer was concerned, they were just generic words, and as a result, thousands of hits were supplied, effectively burying from sight any correct hit on the blog in question.

In another apparent example of search-engine-gone-awry, the same blogger was angered to find not only just re-blogs of the blogger’s original post, but also the many hundreds (thousands? tens of thousands?) of hits that were for blogs which somewhere in their titles or texts had used the word which the blogger had been using for years as a blog title.

In the first place, blog posts get time stamps when they’re published, and because the most recent occurrences of a search term will show up first on a hit list, any re-blogs would appear before one would find the original post. Secondly, getting in the way are the more recent posts of all those thousands of bloggers who, in the interval since the post in question was published, wrote something that used the same word for which the unhappy blogger was searching. Again, if what you’re searching for is a fairly common word, all you’re going to get is rubbish.

My blog post inventory, accessed from my Dashboard, showing my latest 4 posts. Note that the time stamps include hour, minute and second. (Click on this image to enlarge it.)

The question then arose about the “ownership” of the “name” of the blog. The angry blogger, who had paid for a plan, apparently was of the impression that what had been bought was the word that is being used for the title that appears in the header of the person’s blog. The blogger expressed the belief that exclusive rights to use that blog title had been violated, because that “name” had been “bought” but was in use by other people to identify their blogs. This is a grievous misunderstanding.

The “name” that one sets up for a blog (whether free or paid), is not the text that can be put in the blog header: that’s just the title of the blog. In my case, the actual blog “name” is a variant of the domain name possessed by the blog host (WordPress, whose domain name is “wordpress.com”).

Because computers use telephone systems to navigate the internet, a domain name is nothing more than a telephone number. This number appears as the universal resource locator (URL) that shows up in the address bar at the top of a window.

When you set up a free blogsite, what you get for a domain name/URL is essentially the WordPress phone number with an extension number added. When you pay to have a blog account, WordPress obtains a completely different phone number, and rents it to you. (The only thing you own at WordPress is your copyright to the original content you post on your blog, and even that can be contested by pirates and plagiarists, if you haven’t registered your copyright.)

Thus, the title of your blog (and any associated tagline you may choose to add), which you put in the header, is not your domain name, and you don’t own it, no matter how long you’ve used it. Of course, you can choose to reproduce your domain name as the title in the header, but that still doesn’t mean you own, or have the exclusive right to use, the title of your blog. (In fact, according to US copyright law, titles cannot be copyrighted, which is why you can often find several books, fiction and nonfiction, all covering different subject matter, but which have exactly the same title.)

(Click on this image to enlarge it.)

As far as any stupid computer’s search engine is concerned, the title that appears in the header of a blog is just another piece of generic text, and it gets treated just as any other word gets treated in a search: it has to take a number and queue up in a hit list, like any other block of text. Therefore, if you want people to be able to find your blog by searching for its title, you have to make sure that your title is unique enough to have few possible duplicates.

Likewise, if you want your blog to make it to the top of a hit list when a search term matches your domain name/URL, you have to make sure that you pick a suitably unique set of words, or combination of words and numbers. You also need to be sure that your domain name/URL can be easily identified with you, personally, or with something special about your blog. Then, folks will find it.

At this point I would like to report that when I searched for the missing blogger, I used the blogger’s full name, the blog’s full domain name/url, and with Caps Lock on, I entered the common word the blogger uses for a title. I did these separate searches when I was logged in to WordPress, and when I was not, and used two different browsers (Chrome and Firefox). I used the dedicated search engine page at https://en.search.wordpress.com/ and the search engine on the “Reader” page. At the dedicated search engine, the original blog post came up within the top three hits of the hit lists that were returned; at the Reader search engine, only the blog title failed to turn up the desired results, indicating to me that the word is too common. As far as I was concerned, the blogger was not “Lost in Space,” and the search engine(s) not broken.

Now, it’s entirely possible that drag-and-drop experiments by WordPress Unhappiness Engineers have screwed up search engine optimization by introducing strongly graphics-oriented functionality to what is now a heavily text-based application. That kind of messing around is understandable, because WordPress has tons of techies on payroll who have to justify their existence there by fixing things that ain’t broken. The problem with it, is that those of us who are more literate than the point-and-grunt generation may prefer to communicate with real words that can be coded into machine-readable language, rather than to push around pictures composed of what Microsoft PowerPoint called Word Art. Sometimes, WYSIWYG just doesn’t cut the mustard.

The only difficulty I’ve had over the past few weeks are isolated incidents of my userid not working, when I tried to log in. I keep all my usernames and passwords in a spreadsheet file, and all I have to do is copy and paste what I need into a log-in field, so I know that I hadn’t misspelled anything. What I ended up doing to gain access was to temporarily switch to using my e-mail address, and I got in all right. The next time around, my username worked just fine. Go figure.

The really important thing about all this is not to get our knickers in a twist. Turning WordPress Woes into WordPress Wars is as futile as Pickett’s Charge, or the Battle of the Somme: frontal assaults against an entrenched enemy’s strongest point, in an uphill battle.

We can make complaints to the Unhappiness Engineers, and when the changes that are made don’t suit us, we can quietly vote with our feet (if that’s what we type with) and take our blogging elsewhere. We probably won’t be missed by the WordPress Warlords, because there will always be a new generation of clueless cannon-fodder users, who don’t know how good it used to be, to take our place.

In the end, it’s WordPress’s business, so they can run it into the ground, if that’s what they want to do.

* Apologies to Erich Maria Remarque.

Advertisements

7 Comments

Filed under Uncategorized

That’s Artificial Intelligence, For You.*

IRISH FIREBRANDS: A Novel ~ and Other Works by Christine Plouvier, Indie Author

355px-Ernest_Hemingway_in_Milan_1918_retouched_3 Ernie

375px-Edward_George_Earle_Lytton_Bulwer_Lytton,_1st_Baron_Lytton_by_Henry_William_Pickersgill Eddie

Compare these two paragraphs:

On a cold windswept street, this was a warm, cheerful place with a big stove in winter, tables and shelves of books, new books in the window, and photographs on the wall of famous writers both dead and living. The photographs all looked like snapshots and even the dead writers looked as though they had really been alive.
~ Ernest Hemingway, A Moveable Feast

… and …

It was a dark and stormy night; the rain fell in torrents, except at occasional intervals, when it was checked by a violent gust of wind which swept up the streets (for it is in London that our scene lies), rattling along the house-tops, and fiercely agitating the scanty flame of the lamps that struggled against the darkness.
~ Edward Bulwer-Lytton, PaulClifford

My question is, “Exactly how is the first paragraph supposed to be better than…

View original post 632 more words

Leave a comment

Filed under Uncategorized

That’s Artificial Intelligence, For You.*

 

355px-Ernest_Hemingway_in_Milan_1918_retouched_3

Ernie

375px-Edward_George_Earle_Lytton_Bulwer_Lytton,_1st_Baron_Lytton_by_Henry_William_Pickersgill

Eddie

Compare these two paragraphs:

On a cold windswept street, this was a warm, cheerful place with a big stove in winter, tables and shelves of books, new books in the window, and photographs on the wall of famous writers both dead and living. The photographs all looked like snapshots and even the dead writers looked as though they had really been alive.
~ Ernest Hemingway, A Moveable Feast

… and …

It was a dark and stormy night; the rain fell in torrents, except at occasional intervals, when it was checked by a violent gust of wind which swept up the streets (for it is in London that our scene lies), rattling along the house-tops, and fiercely agitating the scanty flame of the lamps that struggled against the darkness.
~ Edward Bulwer-Lytton, Paul Clifford

My question is, “Exactly how is the first paragraph supposed to be better than the second paragraph?” I’m genuinely perplexed about this, because what seems to be merely a matter of personal taste in writing styles has somehow acquired the force of law, among writing gurus and publishing gatekeepers.

The Hemingway excerpt is purportedly from his posthumously published memoir (NB). Papa is supposed to have edited it before he died, but it’s been issued more than once, having passed through various sets of his surviving family members’ hands. I have no idea whether or not this paragraph is pure Ernie.

According to the faceless, nameless contributors to Wikipedia, “[Hemingway’s] economical and understated style had a strong influence on 20th-century fiction.” A long time ago there used to be a “Bad Hemingway” writing contest, and Papa himself is reported to have ridiculed the style of some of his writing. But emotions may run high, here, so in an effort to find unbiased input, I resorted (as many do), to Artificial Intelligence: the eponymous Hemingway App.

Running these writing samples through the program yielded equivocal results. Hemingway received a writing grade level of 15 and a grudging “OK” for quality, but was censured for the first sentence’s being very hard to read, and for one adverb in the second. Apparently Papa forgot to be sufficiently economical and understated to make his namesake App happy.

Bad Bulwer-Lytton writing competitions are still held, so the Baron is routinely hauled forth, wearing a hair shirt and a scarlet letter, to be pelted with rotten aubergines. This time, like Ernie, Eddie got slapped for his having one very hard to read sentence and one adverb, in addition to one use of the passive voice. Oddly enough, the software refused to assign a writing level or a quality evaluation: it only said, “Not enough text.”

How this could be true, when both samples are 58 words long, puzzled me, at first – especially when the passage, “See Dick run. Run, Dick, run!” was assigned Grade 0 and was evaluated “Good.” So I scrounged up a 55-word sentence from Irish Firebrands. Like Bulwer-Lytton’s paragraph, it was accused of “Not enough text,” while also being a very hard to read sentence with one adverb. We all seem to be in good company – Ernie, Eddie, and I.

Then I tried the 6-word “story,” For Sale: baby shoes, never worn, which has an unsubstantiated attribution to Hemingway, and it merely received the reproach, “Not enough text.” By my simply substituting a full stop for the colon, it garnered a Grade 3 and the praise “Good.” Evidently the HA incarnation of AI depends on periods, to help it make sense of the words it sees. The “look-say” people must be proud.

I didn’t try to analyze any of my graduate school nonfiction research papers, but I suspect that the limitations of the Hemingway App make it inappropriate for all but the most simple technical writing; nevertheless, in my online roving I ran across someone (self-identified as a teacher) who had suggested installing the software on school computers. Now, that’s scary.

Does Papa Hemingway deserve the kudos?
Does Baron Bulwer-Lytton deserve the bad rap?
Has independent publishing made these moot points?
What do you think?

* Accent on Artificial.

NB: I found the Hemingway excerpt (without any accompanying commentary) at another blog, possibly posted as an example of “good” writing, because when I posted the Bulwer-Lytton passage to comments with the above question, the blogger chose not to publish my inquiry. That’s okay. It’s a free country. When I went back to the blog to check again, there was one posted response to the blog, and it was favorable towards the Hemingway excerpt. I suspect that the blog already has thousands of loyal followers, anyway, so stimulating a discussion apparently wasn’t the point of that post.

1 Comment

Filed under books, Uncategorized