LilTanker has a better basic understanding of the issues involved, in terms of real overhead and usability.
As a user, if I want to search within the topic, it's a breeze to press ctrl-f, and search the page. I'm guessing that the google ad inserts work a lot better with the whole topic on the page. It sure makes for a bunch more coherency. It also works better with the search engine spider which can just come along and suck it up all at once, and links stay better, and so on.
What happens so often, is that folks install a bloated OOP, e.g. CMS, when all they need is a couple of functions and they would be done. It goes beyond idiotic sometimes when folks think they have to do bloated CMS when a simple database cursor and a for loop is often all that's needed, instead of some big, bloated, 50,000 line piece of junk.
Chris has done a pretty good job in the way that the programming works. The usability is good, even for some of the dumbest users.
There's a search function with more functionality at the top of the forums section, FYI.
The youthful arrogance, and inexperience, of folks like charlitos, shows through brightly. liltanker clearly is the more knowledgeable, since I'm taking sides. liltanker has a much broader understanding of data processing, in general, and speaks in a more qualified way about all the issues that come to bear.
Granted, loading the page all at once isn't' that spiffy on a mobile device, but, long forums don't lend themselves well to those devices anyway. This site is not geared in its business model towards those devices, clearly, nor would it lend itself well to that.
Allowing embedding of off-site objects is, of course, going to slow the page load, but, adds the value of compelling, and not so compelling, multimedia, as well.
At the end of the day, unless you have everyone paying for use of the site, the more possible click thru links that end up on the page at any given time, the more likely there is a revenue stream. A site can't bleed money, day in, and day out, if it expects to stay online.
I can't imagine a worse idea than throwing all the extra overhead of a bloated CMS on this. Dumb, dumb, dumb, in so many ways. Just about nothing could be a worse idea.
One of the wonderful things about custom programming is the flexibility, and the sparseness, it provides.
I've cleaned up CMS messes, and so has Logan, time and again. It's makes good sense, if you're a programmer, to build the stuff custom, and nice and tight, and just the the way you want it, with as many outward faces as you wish, without all the headaches of a big bloated slow piece of software that's a pain in the ass to maintain. Dumb, dumb, dumb.
We've seen CMS systems peddled off to the folks who least need them; who least understand how to use them; and that completely ruin the look and feel of a site.
With regard to database connections and recordset / cursor overhead, liltanker is dead on in his statements. connection management is almost always more costly than a few cycles on the CPU. Text takes little bandwidth, especially on well constructed custom code with style sheets and the like.
Most times, the simplest solution is the best one. Almost always the real cost of ownership (and it IS about business) is lower with the simpler system.
charlitos would do well to study up on how the guts of how I/O systems work, how recordsets / database cursors work, and how TCP/IP works (although, this database may used a named pipe), and how HTTP 1.1 works. What he'll find is that the whole SECTOR is read into memory from disk; that scanning an index has low overhead; that the real overhead is in the HTTP negotiations, and the session setups, as well as the database connection (yes, there a persistent connections sometimes). My point being is that charlitos doesn't realize how little he really knows. True wisdom is knowing that the more you learn, the more you realize how little you truly know. Clearly, charlitos did not take time to think through what he was saying, lacks experience, or didn't do the research required to consider all the factors in the user experience, and in the performance management. charlitos does not have a comprehensive understanding of all the factors that come into play in the delivery of the content.
It's a no brainer.
Like liltanker, I've cut my teeth on some big projects like matchmaker.com, vocalspace.com, audioacrobat.com, byoaudio.com, audioblog.com, hipcast.com, and, most recently, oneseason.com, building highly-scalable, highly-available, high-performance, 5-9s, systems for nearly 2 decades, with over 3 decades in the field. The latest thing we have been doing involves the EC2 Supercomputing Grid. I've worked for GTE/Verizon 3 times, ATT 2 times, and IBM 2 times, on large project implementation as well as startups. In November, I turned down a job as a level 4 enterprise systems admin answering to the director of internal IT at Verizon (I wish I hadn't cuz now we're struggling.) I giggle at some of the kids without a hint of humility, who are clueless about being clueless.
The more you learn, the more you realize how little you truly know.