Questions: A Publishing Strategy

[tedg]

I am developing a database in FMP 12 because the content is rather dynamic and I plan to use many desktop capabilities of FMP. However, the whole idea is to publish it on the web.

This question is about whether my strategy - as described below - is wise.

I plan to master FMP to a degree and build a layout (using Classic theme) that emulates the final web page design as much as possible.

I plan also to build a web page in a Mac program called Freeway, something I know how to do. If this were a PHP/mySQL project, I would need someone to do some heavyweight coding to build elements.

My supposition is that I can find a FM hosting company (probably ODITech.com) and put the FM database up there. A very short term testing period would use IWP. Once I know all is working, I’d ‘connect’ my web template with the database by simple PHP.

This assumes that it is wise to do as much as possible in FMP scripts and calculations rather than in PHP assembly.

For instance, instead of the IWP layout presenting raw fields to be assembled by PHP, I would assemble much larger synthesized blocks. An example: I have film comments that have many components (film title, year, possible foreign title, comment title, comment date, comment text...)

The block of the combined comment presentation with styles would be assembled in the layout, rather than by the PHP server. I hope this is clear.

Is it a wise strategy?

[Fitch]

If you're planning to go to PHP, why bother developing an IWP solution?

[tedg]

Because:

1. It gives me a chance to test that scripts I make on the client work on the server.

2. It allows my eager readers an interim solution while I work on the junky programming model of php/css/html.

3. It enforces a discipline (as described above) of doing all the computations in FMP, and delivering simple strings via PHP.

Meanwhile, I am worried about the “user limit” on FMP. Does this effectively rule out a general purpose data-driven website with expected 1,000 users per day (to start)?

[brian rich]

Hi Tedg

I think you are making a lot of work for yourself. I always considered that the primary purpose of IWP was to make putting a Filemaker database on the web quick and simple - it's not something that you tinker with. I think you will find it very difficult to get IWP to work smoothly within the context of another website. I would go down the CWP route for this kind of work, which means you can control precisely what appears on the website - the downside is that the devopment cycle is three or four times longer. It also sounds like you are building your website from raw HTML, which means many additional hours of development work

I'd suggest that you consider using Wordpress to build your website. Wordpress is an open source package which is offered as an installable application from many Web Hosting companies. Install the application, choose a theme, set up your pages and it's live - a couple of hours work. Wordpress has it's own bunch of plug-ins which allow you to do more or less anything you want with the site, with very little development work. I run a Wordpress website http://awcc.org.uk where I use Filemaker to upload discrete and inline files to the site on a regular basis using FTP. Going down this route means I don't have to have the database hosted on the web anywhere - I send the data from Filemaker, Wordpress does the donkey work. Unless you have a need for real time updates via the web to your Filemaker database, I'd recommend keeping the database local.

If you want the ability to manage Pages, Posts, categories etc directly on your Wordpress website, then check out Lauren Kuhlmans Wordpress Filemaker Sync plug-in see: http://wordpressfilemaker.com/

HTH

Brian

[webko]

2. It allows my eager readers an interim solution while I work on the junky programming model of php/css/html.

I think you're going to give yourself a world of pain - IWP is far more junky than anything that a MAFMP / WAFMP stack could do... It is really *only* for simpler systems unless you want to spend a lot of time hacking it (IMO).

3. It enforces a discipline (as described above) of doing all the computations in FMP, and delivering simple strings via PHP.

And I'm not sure why you would do this either - simpler to deliver, but against most web programming philosophies, where the database is a data repository only, and the logic is exercised at the web end. Even hardcore CMS systems built on MVC principles do the all the logic at the web level...

Meanwhile, I am worried about the “user limit” on FMP. Does this effectively rule out a general purpose data-driven website with expected 1,000 users per day (to start)?

Which user limit? For IWP, it might as the sessions are kept live. For CWP, not even close, as sessions last for as long as it takes to load each page.

Oh, and anything that uses globals at the IWP level will probably not work for CWP either (not totally relevant but always good to know)

[tedg]

I *really* appreciate the guidance. Would like some clarification though.

Regarding the FMP Server limits, you are saying that this is not a problem for a site that could get a few thousand hits a day?

Regarding the IWP solution and the logic-in-the-database issue, let me explain the rational a bit and see if you still recommend the against them.

I will be doing a lot of work on the desktop: building, populating and expanding the DB. I will need most all of the scripts I am creating for this anyway. If the result is useful for IWP without hassle, I’ll use it. If IWP requires tweaking, I’ll just put the energy into the custom solution.

The logic-in-the-database strategy is based on the following assumptions:

• It makes sense to just use the FMP database for the web backend as it is — scripts and all — rather than converting to mySQL.

• I can get pretty close using FMP logic to delivering the exact data-elements that I need for the website. I believe this is within my skill level. I also can create the website using Freeway, having blank areas where the data elements go. This is also something I can do.

• Connecting the dots will be (I imagine) a simple php statement that says “put that here.“

Creating the logic in PHP means getting a consultant who will want to do all the site work. If it is the guy I used before, he will want to transport all the FMP info to mySQL.

However, for me to do what I plan in FMP to deliver web-ready results, there will be complex portal work that I will need help with. Since I am investing in FMP skills, I’d rather do it here if possible rather than learn a second language.

I am basing the design of the data-driven site on an existing static site that will be a sibling:

http://filmsfolded.tedgoranson.com

The design of the data-driven site is attached.

FF Site design.pdf

[tedg]

The advice I got above was good. IWP is a no go. Not even close. Not sure whether the work I did helps with the CWP solution or not.

Will proceed with CWP, but still assembling fields in FMP.

Thanks.