Talk:Z1 (computer)

This article is rated Start-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects:

Computing This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.ComputingWikipedia:WikiProject ComputingTemplate:WikiProject ComputingComputing articles ??? This article has not yet received a rating on the project's importance scale. This article is supported by Early computers task force.

The article lede states "The Z1 was the first freely programmable computer in the world which used Boolean logic", while the Design section states it used floating-point logic as "the Boolean logic unit was not realized". NEJ — Preceding unsigned comment added by 193.44.1.81 (talk) 09:18, 11 March 2014 (UTC)[reply]

Which is it?

—me_and 23:41, 11 December 2013 (UTC)[reply]

I don't see the constradiction because the two uses of "Boolean logic" have different meanings. In the first case it means that the whole machine was built with binary circuits, which are equivalent to both symbolic logic and to Boolean algebra as Claude Shannon proved around this time. In the second case, there were several separate units that implemented the machine's instruction set. The part that handled the floating point instructions (FADD, FSUB, and so on on a modern processor) was actually implemented while the part that would handle the boolean instructions (AND, NOT, OR) was not.Jecel (talk) 23:20, 2 May 2014 (UTC)[reply]

in this edit, User:David spector (who apparently wrote most of this article. Thanks, David), added the following quote:

"The relays [in the data storage memory unit] appear to consist of a number of thin strips of metal lying between two plates of glass, not more than a quarter of an inch apart. Strips running in one direction through this layer represent the 16 cells, each containing one numerical quantity, while those in the perpendicular direction represent the individual digits, etc., of each cell. Motion of the strips is controlled by electromagnetic relays that engage or disengage individual strips with an arm providing mechanical impulses at regular intervals. It would appear that motion of a particular strip representing a chosen cell exposes other moving parts, corresponding to each digit position in the cell, to the motion communicated by the various transverse digit strips. The whole mechanism appeared quite compact and simple; however, detailed examination or description was denied."^[1]

Reading the source, two things are clear: (1) - it was written no earlier than 1943, and most probably after the war, and (2), it describes a machine that existed at the time of it was written.

These two facts, together with inconsistencies between the content of the quote and the Z1: (as described in this article and elsewhere), for instance, Z1 being a 22-bit floating point machine while this describes a 16-bit machine; the quote discusses electromagnetic relays which were not present in the Z1.

These discrepancies make it clear that the "Zuse machine" discussed in the article is a not the Z1. This may be appropriate for another article, but it does not belong in this one.

I do not think it provides useful information about the Z1, and it's confusing, since it presented as if it describes the Z1 itself. So, I'm removing this quote from the article. peace - קיפודנחש (aka kipod) (talk) 17:16, 18 February 2017 (UTC)[reply]

I agree that this description of the unique relays of the Zn machines is confusing. However, I have seen several other descriptions, all much briefer, and all confusing. For example, one description mentions convex cylinders or half-cylinders, yet these are not seen in photographs of these relays. Also, the construction of the relays was said to have improved from one Zn computer to the next. The early relays were sensitive to trolley car movements near the apartment where it was constructed, but the last version of the computer appeared to have relays reliable enough to be used in commercial telephone systems.

I edited the article because it said next to nothing about the relays, yet they are the heart of what made the Zn computers effective.

I believe that all the Zn computers pioneered in implementing floating-point hardware. I don't recall any of them using 16-bit fixed-length words, but my memory might be wrong.

I recommend leaving the article as it is until someone finds better descriptions that are reliable. I have yet to find a reliable description of the Zn computers. Let's not make the article worse. Let's be patient until new documentation is found. David Spector (talk) 14:29, 19 February 2017 (UTC)[reply]

That source (AMS, The Zuse Computer) describes unfinished Z4 (The Origins of Digital Computers: Selected Papers p. 156). — Preceding unsigned comment added by 89.25.210.104 (talk) 20:37, 7 May 2018 (UTC)[reply]

Reading Raul Rojas' paper on the Z1's architecture (https://arxiv.org/abs/1406.1886), the reconstruction is 24-bit, and there's no mention of the original being 22-bit. The reconstruction has 64 words of memory while the original is stated to have had 16. http://zuse.zib.de/z1 says that the original was 22-bit, but also that it had a memory of 64 words, not 16. The site also claims nine instructions taking 1-20 cycles, whereas Rojas claims eight instructions taking 1-21 cycles. What gives? KetchupSalt (talk) 16:01, 28 November 2021 (UTC)[reply]

Yeah, I had noticed this as well, and made both of those fixes (across two edits, as I didn't notice them at the same time). I couldn't find a source that backed up the original claim, and Raul Rojas' paper is rather detailed, so presumably the original Wikipedia entry was in error. MegaZeroX7 (talk) 23:06, 30 August 2022 (UTC)[reply]

Looks good. This leaves the question of 22 vs 24 bits. In most drawings I've seen the memory is arranged in modules of 8 bits each, and this would make sense for saving labour building the thing, rather than building two 8-bit modules and one 6-bit. All we have to go on for 22 is the zib.de page. On the other hand the Rojas paper is mostly about the reconstruction, and the original memory might have been built as just a 16x22 bit thing. Keeping it as 22 at least doesn't overstate its capabilities. KetchupSalt (talk) 23:29, 30 August 2022 (UTC)[reply]