AnalogX

http://analogx.com

radsoft.net receive regular inquiries from visitors who want to know when they'll finally be able to read something about the AnalogX product line, and at last the answer is 'now': you can read about it now.

AnalogX is a phenomenon. A music group of one, a part time programming entity, it has rapidly become one of the most visited sites on the web. One can almost intimate the same development as for CmdrTaco years earlier.

AnalogX gives the impression of 'churning out' the applications, one after one, week after week. An oft-heard comment is: 'that guy has a new app every week - or at least he seems to'.

Others, such as Tina Cusperson at techsightings.com, are not as enthusiatic. Tina has a hard time palating the excess ego exuding from the AnalogX site. In all fairness, Anal has published this review and seems to take it in stride.

But the issue here is not how KEWL AnalogX is, how GREWVY his programs are, but how well they're engineered. The short summation: AnalogX is definitely not a REAL programmer. His background would seem to lack the requisite pedigree, and the resulting executables, while a step above Delphi or VB, are hardly much more. And that's at the very best. At the very worst, despite rather extensive precautions at times, his software can still bring down your operating system - that's right, crash it, and bad.

Make no mistake about it: many of his programs are very useful and many seem to be very well written, and it's all for free (Anal will reap his millions elsewhere as did Malda before him). And make no mistake either: this is not a dissection of Anal's personality as the Cusperson review, nor an appraisal of the site. In fact, this is not even a review of the usefulness of the software, only a look at how well it is engineered.

Anal uses a compression technology to make his images smaller. Which is all fine and good - the problem is that most of these programs, very straight forward in nature, shouldn't need any compression anyway, but should weigh in at 30 - 40KB tops. But all the AnalogX images are around the 200KB line - each and every one.

In fact, because this compression technology (EmuCore) is in use, it's almost impossible to run any diagnostics on the images - not until runtime at least, when the radsoft.net duo of X-tool / Memview can come to the rescue.

A typical Anal image weighs in at 200 - 250KB, uses eight sections in memory (an inordinately high number), but expands to significantly less - most often under 200KB. Scanning this memory reveals the following:

• The compiler and linker come from Watcom, and are dated 1995.
• A static Watcom runtime is linked in.
• Debug info is not stripped.
• Line numbers are not stripped.
• The images claim to use 'low reversed word'.
• The target OS version is '1.0b' (?).
• Images contain about 10KB of uninitialized data (bad).
• Images do not contain their own version information (either fixed file info, string file info, or as part of the standard linker general info).
• Images are targeted for the Windows GUI subsystem version 4.00.

A lot of the above discrepancies may be due to the development tools being rather 'out of date', but the general impression is 'WTF is this', for none of this makes any sense, or is of any advantage to the program at hand.

All AnalogX images are ripe with debugging information. Simply append '-debug' or '-window' to any AnalogX startup and see for yourself. While it is admirable to have a debugging alternative, it is not good for this extra bloat to ship with release executables, as outlined elsewhere at this site. Software is supposed to be adequately debugged before shipment, not afterwards, and end users are not supposed to double as Anal's guinea pigs. Granted, that the sizes of AnalogX images are small compared with Delphi or VB, but that is not the point: packing all this excess code, which can hardly be used by ordinary end users anyway, is just not right. That's not the way things are done in the industry. (By way of comparison, Windows NT Exploder contains only 272 bytes of debugging information - standard for NT system modules - strips both its debug info and its line numbers, and uses an aggressive working set.)

AnalogX images take an eternity to load - but again, if you're used to waiting on Delphi and VB apps to do the same, it might not seem so bad. Part of the reason for this is that the EmuCore compression technology has to expand the files into memory, and part of the reason is that the images are not optimized (optimization and debugging information are mutually exclusive), which also contributes to their seeming 'sluggishness' when running.

And all of this taken together - EmuCore compression, debug info, etc. - might go a long way to explaining why AnalogX programs invariably claim so much real estate on disk.

YMMV, but we receive complaints about AnalogX support. Support is an integral part of any software product. radsoft.net is known not only for its support, but for providing customized solutions of its product line when possible. But what has been reported to radsoft.net would indicate a radically different approach at AnalogX: the software is good as is, needs no corrections, no enhancements, and comments are not welcome.

One AnalogX program, used for monitoring ports, goes absolutely bonkers when the wrong kind of sockets connections are made - it has no recovery logic at all. The end result is that the system has to be taken down. This was reported to AnalogX by a visitor to radsoft.net. The reply was curt and dismissive: the AnalogX program was in use by key US government authorities, there is absolutely nothing wrong with the program, be gone. When AnalogX released an update to this program the other week, our visitor hoped for the best - only to find that the fatal bug remained.

One of the more recent additions to the AnalogX arsenal, TextScan, is considerably below the current level of quality. This program has been reported to crash Windows 9x. It is also horrendously slow, and leaves gaping holes in its user interface. Why things have suddenly deteriorated is not known.

Summary

Is it good stuff? Certainly. Is it quality software engineering? Doubtful. Will one get adequate support for the products, freeware or no? YMMV, but from what has been reported here - no.