Post #426,396
11/19/18 7:04:28 PM
11/19/18 7:04:28 PM
|
Well, it's always something.
I've been quite happy doing my huge clovegarden.com Web Site with very basic HTML4, and I run every page through the W3C verifier so I know they are correct. As of this weekend, the W3C verifier no longer pays attention to the instructions at the top of the page and verifies everything as HTML5 and CSS3, which are radically different from HTML4, and previous, none of which required CSS. So, I've converted one each of the four page formats I use most often, so I can use them as templates. That leaves only a little more than 4000 pages to update. Fortunately, Web Browsers will be paying attention to those instructions at the top of the page for a few more years, at least.
|
Post #426,400
11/19/18 10:49:52 PM
11/19/18 10:49:52 PM
|
That is a pain!
Especially when you consider that nothing is gained for the all important content. Alex
"There is a cult of ignorance in the United States, and there has always been. The strain of anti-intellectualism has been a constant thread winding its way through our political and cultural life, nurtured by the false notion that democracy means that "my ignorance is just as good as your knowledge."
-- Isaac Asimov
|
Post #426,402
11/20/18 9:06:46 AM
11/20/18 9:06:46 AM
|
Have you considered a templating system?
Not runtime, but a static site generator like Hugo or Jekyll. This method protects against the inevitable site-wide changes like what you're describing. Regards,
-scott
Welcome to Rivendell, Mr. Anderson.
|
Post #426,458
11/27/18 9:04:38 PM
11/27/18 9:04:38 PM
|
CSS1/2.1 will do
The higher versions expand on the lower. You only need CSS3 for a web2.0rrhea site where marketing has decided that shiny is better than functional.
HTML5 is similar in relation to HTML4.01. What disappeared are the embedded formatting tags (and unfortunately, you used those rather liberally...)
Overall, the lower CSS specs do improve consistency.
(And as to the browsers still grokking HTML 4.01 Transitional, it is pretty dire. e.g. Edge uses IE7 mode. You're basically chasing rendering bugs in a 12 year old browser.)
|
Post #426,468
11/28/18 11:51:43 AM
11/28/18 11:51:43 AM
|
Re: CSS1/2.1 will do
Actually, HTML5 returned a number of the inline formatting tags that 4.01 had claimed obsolete, including [small] (but not [big>]). I only used "4.01 Transitional", which the W3C verifier was happy to approve with all the inline tags.
As far as CSS is concerned, since I've never used it at all, it might as well be CSS3, which may be final because there is no ongoing effort to produce CSS4. Of course, I'm ignoring all the fancy stuff in CSS3, so I'm probably effectively using CSS1 or CSS2.
Actually, I'm getting along pretty well with HTML5/CSS3, having learned to preserve the appearance of my pages almost perfectly. HTML5 declared a lot of entities "block", and I couldn't live with that, but display:inline in the style sheet fixes that.
One thing that annoys me, is that when I started, I used the [h1], [h2], [h3] tags, but the W3C verifier rejected them as obsolete. Now I'm told their use is critical to good SEO performance. Oh well, I can't use [big] any more, and conversion to [h1,2,3] isn't hard, so I can live with that.
Now that I have constructed examples and .css files it is no more difficult for me to make pages to the new standards than the old.
The problem is 4565 pages to update. Fortunately, I can do a lot by cut, paste and edit from my examples, but it is still a hassle and will take years. I've done a bunch of the most important pages, which tend to be long ones. Updating the three recipe indexes was tedious, but they're done.
Fortunately, Web browsers will continue to display the old format pages correctly for years, but search engines are starting to get a bit picky.
|
Post #426,470
11/28/18 1:42:14 PM
11/28/18 1:42:14 PM
|
If you spec it I'll code it
Given x rules I can probably come up with a conversion script. I'd consider it an honor, your pages are an incredible reference for me.
|
Post #426,505
11/29/18 7:33:26 PM
11/29/18 7:33:26 PM
|
That can be fickle
HTML is notoriously hard to convert wholesale without knowing the structure of the pages. The optional closing tags alone can be great source of fun. Stripping things that can be set on the BODY styling is not too hard, but anything deeper in can be painful.
(But overall, the work is worth it. I have never regretted converting a table based layout to CSS.)
|
Post #426,520
11/29/18 10:38:49 PM
11/29/18 10:38:49 PM
|
I'm aware
I would never attempt a full html parser. But I've done enough screen scraping to hope (which is not a stategy) that his site is predictable enough to do a scripted conversion.
But no response so far so who knows?
|
Post #426,530
11/30/18 11:04:02 AM
11/30/18 11:04:02 AM
|
I have thought about this extensively.
I was thinking perhaps the enlarged picture pages could be automated, but even there there are some that are somewhat different and some that are radically different.
The rest of the pages need extensive work, and their consistency is more an illusion than a fact. While they descend one from another, each is changed in subtle or radical ways to serve an exact purpose.
Besides, I've found it very useful to edit by hand, as it forces me to systematically revisit pages that haven't been touched for 10 to 12 years, and bring them up to standards, mine and the W3C's.
Fortunately, this is a chronic problem, not an acute problem, as Web browsers will correctly interpret the pages probably for around 8 years. Anything new or modified I submit to the W3C Validator will be updated.
In any case, I appreciate the offers, and will keep them in mind. Thanks.
|
Post #426,532
11/30/18 1:20:51 PM
11/30/18 1:20:51 PM
|
Ok, please let me know
If I can handle 80% (random bs number) at least it allows you to focus on the remaining.
Keep in mind non perl programmers are often shocked at what a decent perl programmer can do when dealing with "natural" language. If you can envision it, it is probable that I can code it, no matter how complex you think it is.
|
Post #426,509
11/29/18 7:58:31 PM
11/29/18 7:58:31 PM
|
Strange about the Hn elements
As far as I know, those are not obsolete. There is nothing obvious in the HTML 5.2 standard docs that would indicate they are.
4.01 Transitional was supposed to be a temporary crutch to bridge the gap between the "anything goes" era of HTML 2/3 and the formal 4.01 Strict spec. So, yes, the verifiers will indeed accept pretty much anything. It just wasn't supposed to survive ~15 years ;-)
(And I'll join Crazy: if you need a hand converting or testing, please let me know.)
|
