Advice urgently needed!

Anon

I run a Cosmopolitan-esque fashion website that features a selection of written articles and reviews and so on. Unfortunately, when I designed the site in my early web design days, I made each article a separate HTML page, and naturally it is has gradually grown into an organisational nightmare. Add to that the fact that now I want to modify the design of my site, I've got the prospect of editing each page individually to reflect the new layout and design scheme...!

We learn from our mistakes as they say, so now, as I really do need to re-design the site, I'm thinking that I should do things properly this time - and mySQL and PHP seems the way to go...

I think that putting the reports/articles into a mySQL table and using a handful of PHP scripts to retrieve them is the way forward. That way, all I need to change to alter site layouts in the future should be the PHP scripts. However, the problems begin here: no two articles have the same layout on my current HTML pages, and like a real magazine, a variable amount of photos need to be displayed at appropriate points in the article. Is it wise to do the articles in HTML as I normally would, then paste the necessary HTML code into a "longblob" on a mySQL table? I realise that if the articles didn't have pictures, all I would need do is stick the plain article text in the table and let PHP do the formatting.

Have a look at a typical article from my site:
http://www.houseofcuir.com/apparel/editorials/apparel_editorials_0004.htm

...this is what I need to get into a mySQL table, but I MUST be able to retain "creative control" and have the flexibility to place a variable amount of thumbnails (and links to full-size pictures) in the article, exactly where I want them.

So can anybody tell me what would be the best way of attacking this?

Anon

Rather than storing the images as blobs I've found it easier to just use one column in one of your tables to store the location of the image on the server. That way when you use php to make your HTML you'll just call the info from the location column into the image tag.

As for having each page formatted differently, perhaps you could write a script that takes input such as $align and $bgcolor this would allow you to write html that is reasonably diverse.

Mark153 wrote:

We learn from our mistakes as they say, so now, as I really do need to re-design the site, I'm thinking that I should do things properly this time - and mySQL and PHP seems the way to go...

Anon

Mark, this is a fairly common issue in the development of a content management system.

The common response is to place special tags in the (databased) text that contain object IDs for image references. The tags might look something like this:

<\imgid=4096>

The object IDs refer to separate database records that contain all of the necessary information about the image that is to be embedded -- size, alignment, whether it should be hotlinked to a separate page with a larger version, etc. The image tables might include captions and credits as well.

You then preprocess the retrieved text and convert the image references to proper HTML.

You also could embed the HTML img tags directly in your text, but there's a good reason not to.

At some point you may wish to change how you handle images -- put them on a separate server, for instance, or make a deal with Akamai (which caches images all over the Internet). Or you might want to change how you handle credits and captions. If you refrain from embedding specific information in your content, and instead have those tasks handled algorithmically based on retrieved data, it's a simple matter of changing the embedding code to handle new conditions.

Separating form and content is a good thing.
Let's imagine that you would like to develop another product -- say a special feed for Palm Pilot users -- out of the original data.

You might want to take a look at application server projects such as Midgard (based on PHP) and Zope (based on Python), which are intended to address the very problem you're confronting. Zope is an excellent product with a number of features that you would find compelling, including transactional rollbacks (high-level editorial "undo") and it's free, but it's sort of like converting to an obscure religion.

Anon

Firstly, I agree with (and really like!) your solution. However, I am left a little confused by your explanation! Let me detail to you what I've understood and produced so far, and see if I have this correct:

1) I have created a new table called "images". I have then assumed that the following core fields would be required:-

"imgid" - Self-incrementing number, record identifier.
"thumbpath" - URL of the thumbnail of this image.
"fullpath" - URL of the full-sized image.

...other "nicety" fields such as size, description, author and source of photo would also be appropriate in the finished "product", but obviously not critical at the moment.

2) In my table for "articles", so far I have the following fields:-

"id" - Self-incrementing number, record identifier.
"title" - Name of article.
"author" - Writer of article.
"datetime" - Self-stamping date/time of record-creation.
"content" - The 'longblob' or 'longtext' field (not sure which works best!) for storing the HTML-tagged text of the article.

Now from here, I know that I need a PHP3 script to retrieve and display each record from the table. However, it gets hazy on exactly how I'm going to display the images.

Let's say the first record in the "articles" table has this short, simple piece of HTML in the "content":

So far so good, but when the PHP script for querying "articles" returns a page, all I'm going to get are two lines of plain text, rather than a picture on the second line. How I am going to get the script to turn instances of <\imgid=???> from the table's "content" field into proper HTML <img> tags and then retrieve and display the correct image(s) from the "images" table? (Bearing in mind that there will be a varying amount of images on each different article, and their placement will differ occasionally as well).

FYI, I always pre-size my thumbnail images to the exact dimensions I require before putting them onto the website, so width and height commands would not be a factor, and the image border is, on any article or page, going to be "1".

Any chance of assistance with the PHP code for this please?!? I'm stuck! HELP!

Anon

Mark wrote:
-----------<snip>-----------
...other "nicety" fields such as size, description, author and source of photo would also be appropriate in the finished "product", but obviously not critical at the moment.
-----------<snip>-----------
If at this juncture you already realize that you will add fields, DO NOT WAIT. TRY to think of everything you'll need to manage your data NOW. And put plenty of thought into normalization -- already you may be heading towards having the same information in two separate tables. What if an article's author also supplied the images?

O'Reilly's book "MySQL and mSQL" has a good overview of database design on pp.13-30.

Increase your development time budget on the initial database design and information architecture, even if you have to do it at the expense of visual design work.

Also, check out FastTemplates at:
http://curtisonline.net/theme/phpfast-templates-HOWTO-2.html#ss2.4

Anon

Most definitely agree with you there - I forward-think as much as possible, and will definitely include my-so-called "nicety" fields as well... once I have experimented with the basic "images" table and figured out how its all going to tie-together to produce a finished article!

Anon

I want to echo what Peter Huestis said about the importance of data design. Be sure to read the phpbuilder.com column on database normalization and design:

http://www.phpbuilder.com/columns/barry20000731.php3

You probably should add a column in your article table for "summary." That would be a paragraph summarizing the article that optionally could be displayed below the title on a page that lists multiple articles. Depending on the size of the site you are building, you might wish to add section identities as well to support an organizational scheme.

You also might want to have a table of authors containing name, email, homepage, and other contact information, some of which you might wish to include in the formatted article output and some of which might be useful internally (i.e., phone number). Instead of placing the author's info in the article table, point to the appropriate record in the author table. If the author then changes his name from "The Artist" to "Prince," a single record update handles all necessary changes.

To your question about how to convert <\imgid=???> into the proper html, the usual technique would be to use regular-expression functions. As Peter pointed out, there are several templating packages that you could use instead of rolling your own.

Another way to do it would be to embed PHP function calls in your article -- something like <? placeimage(1234) ?>. Then all you would have to do is call eval() on the retrieved text object.

However, that has a big security drawback: it opens the door to any arbitrary PHP code being embedded into text by an author or editor who is using your system.

Anon

Steve,

Information received and understood about the normalisation and relationship techniques. The link was also helpful, many thanks.

As I said before, I realise that this is the way to do things, but I'm exceedingly hazy on actually putting the techniques into practice. Three months ago, I couldn't even put a simple database together offline. Being genuinely illerate at anything outside of my specialist fields (graphic design and fashion writing), getting to grips to with web design, databasing and PHP is an absolute head-spinner for me! I can just about manage simple, unrelated tables and a minimalist PHP3 script to retrieve from them, but anything more and its over my head!

It's especially hard - not to mention VERY infuriating - when you've got this unique idea for a website (as I know I have), but no clue whatsoever on how to make the thing practical, functional and easy to update/modify. The most disappointing thing of all is that if I do not get something workable and impressive online before the Fall this year, I'm going to miss my window (for the third-year-running) to entice some advertisers from the fashion industry, and I think it will be time to call it a day if I'm not earning a little revenue from the site by the end of November, as my hosting contract runs out then.

As my budget is currently $0 and my "team" is limited to myself, its either rope in some volunteers to help develop some of the PHP/mySQL components (any takers?), or work at it myself and keep asking these painfully obvious questions!

--- (Apologies for the spontaneous sob story...)

I still cannot grasp what you are saying about "function calls" - what I really need is a piece of clear example code or something to that effect, because I cannot make head nor tail of the information on the page that Peter posted.

On a side-note, I've found a site called MediaDepot (www.mediadepot.com) which is, from what I can gather, a free version of Akamai, limited to 100mb of storage space. Might be worth housing my photos and thumbnails there, and then linking across with the (eventual) mySQL "images" table or PHP script.

Anon

OK. GET YOUR HANDS DIRTY. First. Get the book "PHP Essentials" (it's REALLY basic) and work the code samples. Second. Spend at least an hour in this forum per day. Read all posts, even the ones you know will be over your head. Especially those. Third. Put up some functionality on your box. Start with a simple forum like Bare-Bones Bulletin Board. Then go get Phortify from Philip Olson's site after you read Tim's column (and all the comments) here:
http://www.phpbuilder.com/columns/tim20000505.php3
Implement this code on your site. Then you'll be off an running. Believe me.

I started from scratch with PHP in June, and already have a shopping cart, secure login, forums, want ads, and even some limited content management running on one of my sites. WML on another. It's easy, but you've got to work very hard at it. I know exactly how you feel. I'm a converted advertising guy and a photographer who only started with good HTML and Javascript skills. I was really daunted, just like you 10 weeks ago. Now I can hardly believe I've reached some level of accomplishment - due mainly (just like you) to the sheer NECCESSITY of coming up with a BETTER WAY.

Thanks to all in this forum who have helped me!

And best of luck to you, Mark.

Anon

I haven't read "PHP Essentials," just the online manual. But about 12 years ago I learned enough C to be dangerous, and PHP is loosely based on C.

So Mark is a designer and writer, and Peter is a converted ad guy and photographer. Well, I fit into that crowd; I'm a reformed newspaper editor. Resume is on my Web site. I have some time on my hands temporarily, so I've been answering questions here to pay off my karma debt.

Here's the deal on functions, Mark. PHP is a "procedural" language. You can define procedures and give them names, then invoke those procedures later and have them return values (like success, or failure, or a converted string). A procedure that returns a value is called a function, just like in algebra.

You can write a function that takes an object ID as an argument, looks up that object, formats it, and spits it out. Then, in the example I gave awhile back, you could include() that function definition in your Web page, and then invoke it right in the middle of a paragraph of text, like this: <? placeimage(1234) ?>. If that function produces a right-aligned thumbnail of image number 1234, hotlinked to a page containing a full-size version, then Bob's your uncle.

Anon

Peter,

I've successfully implemented the script you referred to on this site - it was relatively painless to get going, so now all I would have to do is integrate page designs around it I guess.

Curiously though (and this might sound a little stoopid on my part!), what use does a logging script have - i.e. what did you use it for? My site is not going to feature any visitor-specific content... although I guess it would be nice to say "Welcome back Mark153!" on the front page! Any chance of a brief explanation of the perks and clever things I might be able to implement with this script, cause I'm not really getting the full-picture, and featuring it still seems pretty benign to me at the moment.

Thanks,

Mark

Anon

I use it (now called Phortify) to authenticate vendors who have control over catalog items and prices, and for users who supply content -- it is also good to have in your forums to keep posters honest.

This script probably taught me more than any other about passing variables between pages, and is a psuedo session manager for PHP3. I understand PHP4 has session management built in, but I can't use PHP4 'cause my host hasn't upgraded yet.

There is a little bit of everything in this library - cookies, updates, inserts, md5, etc.

Another thing you can learn (steal?) from this script is how to send mail with PHP. Useful.

Anon

Yes, keeping posters honest is one of the only logical reasons I thought it might be useful for my site!

I've just installed the latest version of Phorum (relatively painlessly once again I might add!), ready for use on my finished pages, so I would be very curious to know how I could tie this to the login/authenticate script.

Anon

I use a Perl solution for my forums so I can't help you with that. I would recommend another post.

phpbuilder.com uses Phorum, don't they? -- I assume they just get the cookie.