The Lair of the Pythia

When I was first starting out as a programmer I used to make a point of always reading Jerry Pournelle’s column in Byte magazine. Byte has long since bitten the dust, but the Chaos Manor column lives on. I’ve only caught up with it again recently, so I’ve no idea what Jerry has been talking about over the last few years, but in his current mailbag posting he has one of those letters from C programmers: “I would still choose C over Pascal any day of the week.” There’s no elaboration on that, but I’m guessing it would follow the standard pattern of complaining that Pascal doesn’t include a whole host of “essential” features that are not in the Pascal compiler you might see in a college programming course but are included in Delphi. Anyway, Jerry takes the bait, and launches into a resounding defense of Pascal and the philosophy of safe programming in general. You can read it all here (scroll to the bottom).

Everyone else seems to have had their say on the announcement of the sale of CodeGear to Embarcadero Technologies, so I guess I might as well chip in too. By far the most obvious thing about the announcement is that as from now CodeGear and its products will be owned by a company that thought it worthwhile to spend a lot of money on them, not by a company that is get to get rid of them. That has to be a good thing. Also I know where Embarcadero are headquartered, so if they don’t do what I want I can always saunter along to their building and complain. (And then go shopping in the Ferry Building afterwards to cheer myself up.)

On the other hand, as an independent developer, I confess to being somewhat nervous about the fact that the tools I use are now owned by a company that doesn’t have much of a track record of supplying products to people like me, and whose synergies with CodeGear pretty clearly lie in areas that don’t impact me. Whether this will prove a problem or not, time alone will tell, but at least something is finally happening.

Well, I’m upgraded, and so far it hasn’t been too painful.

I decided not to dive straight in because some of my web sites rely heavily on plugins and I wanted to be sure they’d be OK. I’m still not 100% confident, but we have progress.

Priority #1 was Bad Behavior, which the plug-in compatibility list originally listed as problematic. Someone did publish a fix, but I decided to wait for Michael Hampton just in case, and very soon he came through with a new version.

Also a firm requirement on some of my sites was the LiveJournal Cross-Poster. That also has issues, and doesn’t appear to be being maintained by the original author. Thankfully the fix posted here appears to work. However, this does raise a wider issue about upgrading. Generally I upgrade by simply replacing files, but it is clear from that issue that sometimes whole files of WordPress code get deprecated. It would be nice to know what they are so we can remove them. (Yes, I know, I should research it myself and post it to the Codex.)

The only other issue I have had so far relates to a couple of sites that use derivatives of Sadish Bala’s Mistylook theme. Under IE 6 the sidebar drops down to the bottom of the page when you upgrade to 2.5. Other browsers (including IE 7) appear to be fine. Shaving a few pixels off the width of the sidebar div fixed the problem, though I did have one bizarre moment when I had to take off one more pixel because what worked on my test server did not work with the same browser when I sent it live.

(By the way, the browser archive at evolt.org is a wonderful resource when you have problems like this.)

I haven’t yet had a chance to check any of the new features of 2.5, so there is likely to be another post in a day or two, but I do know that not having the categories list easily available on the right is a major annoyance right now. I guess I’ll get used to it, but I foresee a lot more posts in which I forget to set any categories.

One of the great things about running Wordpress-based web sites is that when you want to do something different you can often find exactly what you want thanks to someone’s neat plugin. I’ve written a few myself by now, though nothing that is worth issuing publicly as they are all pretty specialized. However, I’ve also make good use of other people’s work, and it is about time I gave some public thanks, so here goes:

• All of my blogs rely on Michael Hampton’s fabulous Bad Behavior for protection from spam;
• Many of the people I interact with won’t use any form other blogging other than LiveJournal, so I am indebted to Evan Broder, whose LiveJournal Cross-poster works perfectly and does everything I want; and
• One of my sites is making liberal use of Lester Chan’s WP-Polls, and that is also doing a fine job.

Thank you, guys. You’ve been a great help.

One of the less elegant things about open source software is the upgrade process. Because systems like Wordpress are so dependent on third party themes and plugins, it is inevitable that any “breaking” release is going to cause havoc. And any problems are likely to be exaccerbated by the fact that so many of the users have little or no programming skills.

Wordpress 2.3 looks like it has suffered in this way. In the first few days after it came out there was a rash of complaints about broken plugins and themes. A lot of these were apparently due to people making raw SQL calls in themes, which rather boggled me, but again what can you expect with so many inexperienced developers involved.

Plugins had problems too, because 2.3 changed the database schema (surely it should have been called 3.0 in that case?). Those that still have active support have now mostly been updated, but there do appear to be a few weird things still happening.
Read the rest of this entry »

Well, things have been a little quiet here of late, so perhaps it is time for a change of direction.

If anyone is still reading this they may recall that I started this blog with the intention of documenting my experiences with the Delphi Turbos. That hasn’t happened, for a variety of reasons, including:

• The Turbos have proved something of a dead end;
• I can’t talk much here about work I do for clients; and
• I still can’t find a good reason to upgrade from Delphi 7

On the other hand, there is interesting stuff happening. I am using Delphi for PHP, though not the VCL as all of my PHP work is Wordpress-related and Wordpress is not a VCL application. I’ve also downloaded the 3rdRail trial, and hope to find time to take a look at that.

So it would seem that the thing to do is to talk more about general programming issues to make up for the lack of talk about Delphi. My guess is that mostly I’ll be talking about Wordpress. I also suspect I’ll end up saying a few dumb things because I know PHP much less well than I know Object Pascal, but so it goes.

Time for a Wordpress post, I think.

This week’s topic for discussion around Delphi-related blogs will undoubtedly be the newly released roadmap. Most of this will doubtless center around technical features and will be of the “I want my new toys now (and other people can wait for theirs)” type, which is entirely understandable because everyone out there trying to make money out of software wants the best tools for their projects now. For me, the bleeding edge is not a great issue. The only development I’m really interested in is 64-bit, and I’m happy to wait on that while the hardware and O/S implementations sort themselves out.

What I am happy about, however, is the existence of a roadmap. A lot of people have tended to scoff and CodeGear’s worries about Sarbanes-Oxley compliance, and have even suggested that this is a smokescreen for not wanting to say anything. Those of us who have worked in regulatory economics know better. These days it is a really bad idea for a software company to go around making airy promises about new versions of their products unless they are pretty sure that they can deliver. So the fact that CodeGear has issued a roadmap suggests that they, their auditors, their lawyers and whatever consultants they might be using are all happy that it is safe to say the things they said. And that, in turn, makes Delphi’s future look a whole lot more safe and secure. (And if you have clients who are convinced that CodeGear will be out of business soon that’s a very good thing.)

One other thing worth noting is that is appears from what the roadmap says that Studio 2007 will have an upgrade path from Delphi 2007, and not be a separate product. It doesn’t exactly say that, and the small print warns about not using the roadmap as a basis for licensing decisions, but it does seem to imply it, which is encouraging.

Yeah, I know, I’ve been quiet for a long time. I’ve been working, earning money, OK?

But I have finally found time to start having a play with Delphi for PHP. It isn’t much of a start, but it is a start.

I’ve been doing a bit of Wordpress work – essentially adding a whole lot of extra database tables to a Wordpress blog. Now Wordpress is not a Delphi for PHP app, so you can’t just dive in there and use the VCL. However, I have been using the IDE, because if I’m writing PHP the source code highlighting stops me making a whole lot of stupid mistakes that I always used to make in Dreamweaver. For this I am very grateful.

And there might be more. Wordpress plug-ins should not, I think, be written using the VCL. But if you have a bunch of extra tables that are largely independent of the Wordpress data… Now the choice is between writing moderately complex data entry screens that are plugins for the Wordpress admin system, or writing a separate program using the VCL. Wordpress is, after all, just a mySQL database, so as long as you don’t touch the core Wordpress tables it should be OK, right?

Well, doubtless it won’t be that easy. It never is. But at least I have a potential trial project.

I’ve steered clear of the debate over CodeGear announcing a new CEO because I don’t see any value in speculation and conspiracy theories. However, Jim Douglas has now started blogging and his first post contains some encouraging comments. I particularly like the fact that he mentions the “first experience”. Buying and installing CodeGear products can be rather depressing, and leave you wondering if the product itself is as bad as the buy/install/register process. This can be unfair, because the guys who write Delphi products are not necessarily the same guys who run the shop. Furthermore, if the shop and registration process are still managed by Borland rather than by CodeGear staff (or by third parties contracted to Borland) then it may need someone at Jim’s level in the company to get things put right. Here’s hoping Jim’s focus on this area bears fruit.

Gabr has wise words to say about enabling range checking.