So that's what happened to eSpeak
Starving, hysterical, naked survivors tell tales of bloat
Survivors of HP's now-deceased eSpeak middleware have crawled ashore, telling tales of gruesome horror. Or at least they're at least a bit fed up with us praising it so generously, and here's where they get to tell their side of the tale.
The comments - written with the grace of an angel - also have a ring of universal truth: they may remind some of you of your own software bloat nightmares, monsters that collapsed under their own weight, like really fat children.
Here's one survivor's account.
You wrote: "HP's software portfolio is very highly regarded, and in Chai and eSpeak - two you don't hear so much about - it has a couple of crown jewels."
I think I can confidently state that this particular jewel was constructed largely of paste and food coloring.
Imagine a Death Star sized concentration of Java whose density was such that space, time and sanity crumbled in its very presence. Swarming teams of engineers working around the clock without dissent lest the dark Lord himself freeze them with his steely Voice Mail Of Death.
I always found great sport in asking these questions when I was particularly grumpy after a period of unpaid overtime:
(a) What is e-speak? No, really, what is it supposed to actually do? And how? And why? I've read the code and the marketing and I still don't get it. The best answers I heard were usually sweeping generalities encompassing the trivial, impossible, incomprehensible or the jaw droppingly ill concieved.
(b) How many people who've liked it have actually understood it?
(c) Has anyone that's seen the code (without the special injections of course) not turned into a pillar of salt?
(d) If Microsoft thought it would ever work, wouldn't they Nuke it from orbit? (they did conspicuously take aim, and rumor had it they bought the domain when the brand name leaked out).
But the analysts saw HP developing Java/XML middleware, and assumed it was a J2EE killer because the slideware had all the right TLAs in lots of little boxes connected by lines (1 pixel thick) labelled "secure". It was during e-speaks time in the sun that .NET was announced, but I'm sure that was a coincidence...
Various startups announced they were going to use it and were thus guaranteed some press inches and wealth beyond measure.
It only got embarrassing once the betas started shipping. Boot times measured in minutes, MTBFs of seconds, and documentation that brought on an illness resembling haemorragic fever.
But the whole vortex had become self fuelling by then. Externally customers couldn't criticize it, since it would be an admission that their venture was on the rocks. Internally Lord Vader (he with the Whim of Iron) had a crushing grip on his own Storm Troopers so Empress Carly never knew of the fallibilities. Even though it was open source and people did actually download, I don't think we ever heard from any of them again.
The whole thing was finally "scaled down". I believe the phrase used was that it would continue to be "funded in-line with its revenue".
I've been quite amazed that nobody saw the plasma burst of the final implosion and now it's like it never existed. Evidence of it is as hard to find as Nessie or Microsofts WinCE JVM."
Isn't that beautiful?
But how fair? Well, since the sun is shining on this beautiful morning, the sky is blue, and human beings once again look like things emitting billions of fascinating signals, it seems to churlish to leave it like that.
So went to rummage through the plasma. And found this. We'll spare you a 1.1MB PDF download, because we've summarized this document into a simple aphorism, that can be printed on ice lolly sticks and corn dog handles, and distributed to programmers in distress.
Q. How many former COBOL programmers does it take to write an XML interoperability specification?
(Now make the sound of an ice lolly stick turning over…)
A. You can never have enough former COBOL programmers to write an XML interoperability specification. ®
Sponsored: RAID: End of an era?